Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized

May 1975
World Bank
Sector Policy Paper
|
HFLE
J
7C,pyI
LANDREFORM
I

A
 LAND REFORM
CONTENTS
Page
Introduction .......................................... 3
Summary.......................................... 5
Chapter 1: Characteristicsof Land Reform............................. 15
Man and Land........................................... 15
Context of Land Reform.......................................... 16
Dimensionsof Land Reform....................................... 20
Chapter 2: Land Reform and EconomicDevelopment ...... ............. 25
Implications for Productivity....................................... 27
Land Reform and Employment........... ......................... 29
Land Reform and Equity ......................................... 30
Effects on Marketed Surplusand Savings.......... .................. 31
Tenancy Reform...................... .................... 34
Implementation Issues.......................................... 35
Chapter 3: The World Bank and Land Reform.......................... 38
Changing Concerns.......................................... 38
Technical Assistance...................... .................... 38
LendingOperations.... ...................................... 40
Major Policy Options.......................................... 46

Annexes
1. The Context of Land Reform.................................... 49
Ratios of Population to Land.................................. 49
Population and Production.................................... 52
Distribution of Land........................................ 55
Tenantsand FarmLaborers.................................... 60
LandlessWorkers .......................................... 62
2. Experienceswith Land Reform ................ .................. 64
Republic of China.......................................... 64
Republicof Korea.......................................... 64
Japan ........ . 65
India...... 66
Iran...... 67
Morocco ........ 68
Yugoslavia...... 70
Kenya...... 71
Mexico...... 71
Peru...... 72
4L - t'
6
 INTRODUCTION
Land reform is concernedwith changingthe institutional structure
governingman'srelationshipwith the land. At present,the livelihood
of more than half of mankind dependsdirectly on agriculture. Nine-
tenths of this total agriculturalpopulation is in the developingcoun-
tries, where questionsof accessand rights to land are of paramount
interestto morethan 2,000million people.
Land is one of the basicfactors of production for food and other
agricultural products.With food production rising in the developing
countries at about the same rate as population, there is growing
pressureon land resourcesto increaseoutput. Much of this increase
will haveto come from higher output per hectare.Changingthe pat-
tern of landownershipand redistributing land can contribute to in-
creasesin output in somecountries but will makelittle difference in
others.
Conditions governing agriculture vary enormously in developing
countries.But one characteristicthat is common to all is a very rapid
growth in rural population. Thus, while pressureon the land is in-
creasing,the averageman-land ratio is worsening.At the sametime,
nonagriculturalemploymentopportunities are not expandingrapidly
enough to provide adequateincomesfor all those enteringthe labor
market.Somecountries haveprospectsfor expandingthe frontier of
cultivation to absorbmore labor. In other countries,more labor could
be employed in the rural sector through a redistribution of land,
while in yet others changingthe rights to land will makelittle direct
contribution toward absorbingmore labor.
Distribution of landin terms of sizeof holdingsvariesfrom country
to country.The greatestdisparitiesarefound in LatinAmerica.Where
the pattern of land control is skewed, the distribution of income is
generally uneven,although to someextent it is the poorer land that
makesup the larger holdings. In Asia and the Middle East,maldis-
tribution is reflected in the landlord-tenantproblem; the population
is moreevenlyspread,but rightsof accessto landare restricted.Much
of Africa presentsa different problem, as the traditional pattern of
group ownershipand communalrightsiseroded in favorof individual
ownershipwith varyingdegreesof equality.
In terms of land reform policy, therefore,one is confrontedwith a
rangeof cultural and political situations-based on different patterns
of social organization and customs-and with different levels of
development.As shown in Chapter1, at least six land-tenuresitua-
tions can be delineated.The differencesamong thesetypes point to
the varying reforms necessaryto achievemore equitable land access

3
and improvedproductivity in specificcountrysituations.Accordingly,
while it is possibleto identify the need for land reform, it is difficult
to makegeneralprescriptionswith regardto the form of landholding
or pattern of distribution necessaryto achievethe multipurpose ob-
jectivesof development.
Further,one is dealing with a dynamicsituation,where rural popu-
lation growth and changing technology interact with the existing
institutional structures of rural society. The manifestationsof this
interaction are seldom benign for the majority of the land-based
population. A situation that hasseemedrelatively stable and equit-
able for decadescan become untenable.This dynamismmeansthat
a solution which was appropriateten yearsago maybe inappropriate
today. Not surprisingly,therefore, many developing countries are
experimenting with a variety of possible solutions-with different
forms of rural organizations, ranging from communes to private
ownership.
While recognizingthe broad context of the land reform issue,this
paper focuseson a much narrower aspect-the appropriate role of
the World Bank.' In pursuing this question, Chapter 1 looks at the
characteristicsof land reform in terms of both its rural context and
its component elements.Chapter2 examinesthe economic implica-
tions of land reform in relation to the goalsof development.Chapter
3 reviewsthe Bank'spolicy in relation to land reform.Thequantitative
backgroundto land reform in terms of population patternsand land
distribution is outlined in Annex1, while someexperienceswith land
reform programsare summarizedin Annex 2. The policy guidelines
are presentedat the end of the Summary.
'All references to the World Bank in this paper are to be deemed to refer also to the
International Development Association, unless the context requires otherwise. The fiscal
year (FY)of the two institutions runs from July 1 to June 30.

4
 SUMMARY
Landreform involvesintervention in the prevailing pattern of land-
ownership, control and usagein order to change the structure of
holdings, improve land productivity and broaden the distribution of
benefits. In practice, land reform is pursuedin responseto political
pressuresfor socioeconomicchangearising from factors such as in-
creasedpopulation, pressureon a limited land baseor an ideologyof
egalitarianismbasedon more even distribution of land or income.
Land reform, by its very context, has interlinked political, economic
and social dimensionswhich in turn havesignificantimplications for
development.
The systemsof land control in developing countriescan be classi-
fied into six types,as presentedin Chapter1, although in manycoun-
tries examplescan be found of more than one type. Three of the six
types are found in a traditional context: the feudalisticlandlord and
tenant system of some Asian countries; the feudal Latin American
systemof large farms; and the communal landownershippatternsof
manytribal groups (especiallyin Africa). The other three major types
have a modern context: the private ownership of land common in
most marketeconomies;the stateor collectiveownership of socialist
countries; and the plantation or ranch type, which is often inter-
spersedwith otherforms of tenure.
Land reform necessarilyimplies many different kinds of adjust-
ments in an array of situations where there are great variations in
individual equity and agricultural productivity. In most instances,
social or equity considerationsare the main concerns.Thus, when
there are exploitative landlord-tenantsystemsof the Asian or Latin
American feudal type, reform incorporates changesin the rights of
tenants, redistribution of ownership to existing tenants, or the re-
placement of the landlord by the tribe or the community. When
individual ownershipof the market economytype isthe norm but the
' distribution of land is skewed, reform may require subdivision of
large holdings or transfer to the state. In contrast, reform in states
with extensivegovernmentcontrol may involve the transferof some
l landfrom the stateto individuals.
Other variationsof land reform focusmore on the economicuseof
resourcesthan on equity. Where holdingsarefragmented,an appro-
priate reform might involveconsolidationof holdingswithout change
in the patterns of ownership of land. Where communal lands are
eroded or depleted, the appropriate reform might involve a program
of supervisedcooperative land managementwithout changing the
distribution of land. Elsewhere,land reform might involve changing

5
tenancyarrangementswith emphasison providing securityof tenure
so as to encourageon-farm investment.Again, thesedo not require
redistribution but eventually lead to a more economic use of re-
sources.
The typology outlined in Chapter 1 makesit clear that there are
situationswhere land reform is a necessaryprecondition for modify-
ingthe structureof a societyand raisingagriculturaloutput. However,
while land reform in itself may be necessary,it alone is not sufficient
for improving land productivity and distribution of income. Changes
in patternsof landownershipwill not automaticallyleadto an increase
in output or technological change in agriculture. These will come
about only if adequateprovision is madefor the supplyof necessary
inputs and mandatory servicesto the usersof the land. Indeed, as
stressedin Chapter 2, the organization of the supply of inputs to
accompanyany land reform program is essential,especiallywhere
the processof reform leadsto a breakdownof the institutional struc-
ture of agricultureand leavesnothing in its place.
Finally, it must be recognizedthat a policy for land reform for a
given situation cannot be statedin simple terms. Any policy involves
fundamentaljudgmentsabout the adequacyof an existingsystemand
the most appropriate alternative.The judgments of policy makers
differ. The casestudiesin Annex 2 showthat reform-minded govern-
ments,suchas in Kenyaand Peru,havepursueddifferent approaches.
Somegovernmentsfavor individual ownership of land; others favor
communal or collective control over land. Clearly, the policies fol-
lowed are not a matter of economicsalone. Theyalso reflect politics
and ideology, and reach far beyond any purely economic calculus.

Distribution of Landand Income

Although few data are available,the distribution of landownership
is known to be skewed,the degreeof concentrationvaryingwith the
typesof tenuresituation. TheAsian and LatinAmericanfeudal types,
and the plantation ranch types, have high degreesof property con-
centration. The socialist and traditional communal types have low
concentrations.The market economy type falls somewhere in be-
tween. Individual countries are classifiedon the basisof landowner-
ship concentrationin Annex 1,Table 1:9.
Thedistribution of landby size of holding ishighly skewedthrough-
out the world. As shown in Annex 1, Table 1:6, an estimated80% of
all holdings are lessthan five hectaresin size, with about 40% less
than one hectare.Theseholdings account for approximately20% of
all cultivated land, and only 7% of all land in holdings. Considered

6
separately,the pattern in Latin America is particularly skewed.Less
than 20% of holdings(thoseover 50 hectares)account for over 90%
of the total area in holdings,and more than one-third of all holdings
(those less than five hectares)account for only 1% of the area held
(seeAnnex 1, Table 1:8). In Asia, by contrast,40% of the land (ac-
counting for almost 80% of holdings) is in holdings of lessthan five
hectares.
The distribution of holdings by size is frequently usedas a first ap-
proximation in estimatingthe distribution of wealth and income in
the agricultural sector.The skewnessof the distribution of holdings,
however, does not reflect precisely the patterns of distribution of
wealth or income.Thisis because,firstly,all landis not homogeneous;
a concentration of large holdings in a semiarid region may reflect a
smaller concentrationof wealth than a concentrationof small hold-
ings in an irrigated area. Secondly,the distribution of holdings by
size is not the sameas the distribution of ownershipof land; in gen-
eral, there is a greaterconcentrationof landownershipthan of hold-
ings,as evidencedby widespreadtenancy,especiallyin partsof Asia
(seeAnnex1).The distribution of income in theseregionswill depend
largely on the contractualarrangementsbetweenowners and tenants
or sharecroppers.But, in most cases,the distribution of income will
be more skewedthan the pattern of holdings.Frequently,the income
of sharecroppersand tenants may be little different from that of
landlesslabor.

Socialand EconomicIssues
The rural population in developingcountriescontinuesto increase
by more than 2% per year, adding to the already heavypopulation
pressureon the land. Exceptin a few places,there is no virgin culti-
vable land left, so that absorption of more people into agricultural
activity requires more intensive cultivation of land already in use.
The need to absorb more people in the rural areas differs among
developing countries. In many, massiverural underemploymentis
accompaniedby high ratesof open unemploymentin the cities and
growing inequality in the overall distribution of income. Where the
problems are most acute-as in parts of Asia-the emergenceof
large numbers of landlesslaborers in rural areassuggeststhat the
family farm systemasa meansof spreadingwork amongfamily mem-
bers maybe breakingdown.
The extremepoverty of manywho live on the land,and the increas-
ing pressureon the land through population growth highlight the
double challengeof rural development:to raiseproductivity and in-

7
come in agriculture and, at the sametime, to provide more employ-
ment. Accessto land,and the conditions that governaccess,are ques-
tions of major importance in these circumstances.Where land is
marketable, increasingpopulation pressurewill inevitably drive up
the price of land, thus benefiting those who own land. Where land-
ownership is skewed, this will tend to exacerbate inequalities in
income distribution.
These same circumstances(relating to employment and income
distribution) give rise to questionsabout the efficiency of land use
under existing arrangements.For various reasons,landownersoften
prefer to underutilize land, either by working it themselveson an
extensivebasisinstead of through tenantson an intensive basis,or
by leaving it unused.In other cases,tenancyarrangementsare such
that landlords are discouragedfrom making investmentsand tenants
from applying variable inputs, becausehalf the benefits will go to
the other party. In some situations, the fragmentation of holdings
causesgreat inefficienciesin land useassociatedwith transportation,
irrigation and mechanizedoperations(evenon a small scale).In gen-
eral terms, increasesin the population of working age create addi-
tional demands for work and income. At the same time, however,
the additional labor available, if used productively, could serve to
augmentoutput. A strong casecan be madefor land reform (includ-
ing tenancyreform and consolidation)in situationswhere landwould
otherwisebe underutilized in termsof its production potential.
Evidenceon the effectsof changingfarm size (examinedin Chapter
2) indicatesthat the productivity of land-defined asyield per hectare
-is generally higher on smaller holdings than on larger holdings.
The main reason is that smaller holdings are worked with bigger in-
puts of labor than arelargeholdings.The economicbenefits,however,
often depend on the effectivenessof new technology when usedon
small as comparedwith large farms; mere redistribution of land may
not sufficeto raisefarmeroutput substantiallywithout accompanying
agrarianreformsand new services.
Theseeffectson output maybe reinforcedby someof the possible
side effects following land reform. Smallholderstend to consume
more of their own produce and, therefore, market less, per unit of
output, than do large farmers; this may necessitatefood imports to
meet the needsof urban consumers.On the other hand, the addi-
tional food consumedby small farm families might have otherwise
been purchasedif membersof the family had moved to the city. The
consumption of food by poor growers may also be lesscostly than
the consumption of imported or capital-intensiveconsumer goods
by the better-off farmers.Smallfarmersmayalsosavelessper unit of

8
income. The evidence suggests,however, that small farmers save
proportionately more than urban dwellers, and that in the aggregate
they may also have larger savingsthan large farmers,though these
maybe directly investedin the smallholding.
A program basedon the prescriptionthat "the benefits should go
to those who till the soil" is often reasonablein an agrariansociety.
But in a partly urbanizedsetting, thosewho do not work on the land
still require and should havesomerights of accessto the products of
the land. The food and fiber needs(and the spatial requirements)of
the nonfarm population are not infrequentlyoverlookedby the advo-
catesof land reform. In this respect,attention should be paid to both
a minimum and maximum farm size. Thesesizesmight be designed,
firstly, to ensurethat smallholdingsare large enoughto provide food
sufficient to meetwith a highdegreeof certaintythe minimum physio-
logical needs of the farm family; and, secondly, to ensure a scale
large enoughto provide a salablesurplus to meet the needsof urban
consumers,especiallyfor fresh produce. Few land reform programs
provide for sucha minimum limit despiteevidence,from manyareas,
that allowing farmsto becometoo small (relativeto the bestavailable
technology) may be just as unsatisfactoryin terms of equity and
efficiencyasan uncontrolledtenancysituation.

Recent Experience with Land Reform

Experiencewith land reform in the past points to the overriding
importanceof the political factor in securingmeaningfulchange.The
concentrationof control over land provides a power basefor many
groups in developing countries.Land is a symbol of authority and a
source of political power, especiallywhere the landowner controls
the accessof peasantsto their only sourceof security-land. A mean-
ingful land reform programwill inevitably destroyor limit the power
baseof manypersons.It is not surprising,therefore,that land reform
is often a central issuein political debates,and that thesedebatesare
often couched in terms of redistributing political power as well as
wealth. Ambitious programs of land reform will seldom be imple-
mentedunlessthere are shiftsin political sentimentand power. Many
countries have legislated land reform, but only a few can be said to
have implemented it. And in these casesthe reforms were imple-
mented only when there was a change in government in circum-
stancesthat favoreddrasticchange,asin the Republicof China,Japan,
Kenyaand Mexico.
A second factor of importance in making reform effective is the
creation of institutionsto implementthe reformsonce legislated,and

9
to pressfor continuing development. This has usually involved or-
ganizingthe beneficiariesto create follow-up pressure.For example,
in Japan,Taiwan and Venezuelasuitableinstitutionswere established
to ensurethat land was indeed transferred.In other countries,a com-
munity of interestsbetween landownersand officials,combinedwith
an absenceof organizedpressurefrom the beneficiaries,largely nulli-
fied positive reform efforts. The land reform experiencein much of
Asia and LatinAmerica suggeststhat someform of rural organization,
especially involving local representation,may be a critical condition
forsuccessfullandreform.
A third conclusionis that land reform is rarely undertakenwithout
considerableupheavaland lossof production, although there is evi-
dence to suggestthat these costscan be kept small and temporary.
The restructuring of landholdings is often accompanied by the de-
struction of traditional deliverysystemsfor input needsand marketing,
since thesesystemsare almost alwaystied to the operations of the
larger farmers who are dispossessed.Becauseof this, rather than
becauseof any deficiency inherent in the small relative to the larger
farmers, land reform hasoften proved costly in terms of lost output.
Minimizing such costs necessitatesthe provision of servicesconcur-
rently with reform implementation, incorporating as much forward
planning asfeasible.
A fourth considerationrelatesto the problem of perspective,over
time, in assessingthe effects of land reform. As the country experi-
encessummarizedin Annex2 reveal,the effectivenessof land reform
may be relatively limited in the short run, and manysocioeconomic
benefits, such as are associatedwith greatersocial mobility and im-
proved political stability, emerge only in the longer run and accrue
for many years subsequently.The casesof Japanand Mexico are
particularly significant in this respect. While the direct short-run
effects of the land reforms in these countries havenot been consid-
ered wholly beneficial, there is little doubt that the long-run effects
for their total societies have been overwhelmingly favorable, con-
tributing substantiallyto the ultimate economicdevelopmentof both
countries. -

The World Bankand Land Reform

The World Bank has taken an active interest in land reform on
a number of occasions.Concern hasusually been focused on new
or improved possibilities for production following changes in the
tenure situation,with emphasison securityof tenure beinga particu-
larly important theme. More recently, the extent and gravity of the

10
 employment problems and income disparities in developing coun-
tries have causeda new concern over land reform, from an equity
aswell asa productivity standpoint.
The Bank'sexperiencethrough project financing of land reform
hasbeen very limited. In part, this may be becausethere have been
relatively few casesof land reform, particularly in areaswhere the
political situation was reasonablystable and otherwise conducive
to World Bank involvement. But also relevant is the fact that the
financial requirementsof land reform tend to be relatively limited.
Even where the land transferred is purchased from the previous
owners, the amounts involved are usually small, especially where
paymentsare in the form of bonds. In addition, such paymentsusu-
ally constitutean internal transfer(unlessforeign ownersare involved)
and, thus,are not attractivefor externalfinancing.Someexamplesof
World Bankinvolvementin land reform programs,notably in Malawi
and Tunisia,are discussedin Chapter 3.
In general, this report concludes that land reform is con-
sistentwith the developmentobjectivesof increasingoutput, improv-
ing income distribution and expanding employment, and that the
World Bankshould support reforms that are consistentwith these
goals.However, it is recognizedthat the Bankcannot force structural
change; it can only support appropriate efforts within existing
structures. Although the Bank's direct action must be limited, its
preferencesregarding national policy choices and those which are
consideredconsistentwith the Bank'sdevelopmentgoalsare set out
below as country guidelines.Thesesameconclusionsare reflectedin
the subsequentBankpolicy guidelines.

CountryGuidelines
1. Governmentswhich accept a basiccommitment to land reform
5 should consider three components: (i) redistribution of landowner-
ship to reduce the presentmaldistribution; (ii) tenancy reform; and
(iii) consolidation,where necessary.
2. A commitment to land reform implies simultaneousaction to
create or develop an input supply systemto meet the special needs
of the beneficiariesof land reform. This may require either the crea-
tion of new institutions,or specialbranchesor fund allocationswithin
existing organizationsto supply credit, inputs and technical services,
including researchand extension.
3. In sparselypopulated regionsor countries,specially structured
settlementschemescan serveas second-bestsubstitutesfor, or sup-
plements to, the redistribution of land currently in use.

11
 4. It should be recognizedthat a small farm structurecan generate
employment to absorb underemployed labor in crowded regions
where there is no short-term prospect of absorbingit in nonfarm or
largefarm employment.With a seed-water-fertilizertechnologynow
availablethat is neutral to scale,sucha structurecanproduceat least
as muchper unit of landasa largefarm structure.
5. Equity-oriented land reform should be so programmed that
(i) the effectiveceiling on size of holdingsis low; (ii) the beneficiaries
belong to the poorest group; (iii) the extensionand (nonland) input
distribution systemfavorsthe beneficiaries;and (iv) owned and self-
operated land, as well as leasedland, is redistributed.
6. Where efficient large-scaleplantations or ranchesexist, these
need not be broken up, but it should be acceptedthat in such cases
the objectives of reform can only be realized if the enterprisesare
coveredby a progressivetax systemand the workers participateade-
quately in the benefitsof the enterprise.
7. Researchshould be organized to evolve a low-cost settlement
policy. Wherever settlement policy is used to supplement land
reform,settlementschemesshould be plannedto haveapproximately
the same effects as the redistribution of existing holdings. These
effects can accrue if (i) the settlersare the really poor small farmers
or landlessworkersand an input supplysystemis availableto support
their operations; (ii) the size distribution of the new holdingsis equi-
table; and (iii) tenancyis discouraged,and allowed only under speci-
fied typesof contracts.
8. Where the shortageof land is so acute that even with a low
ceiling both smallholdersand landlessworkers cannot be given mini-
mum holdings, preference should be given to smallholdersin the
allotment of land, and a rural works program should be organized
for the landless.
9. Experiencein EastAsian and some Latin American countries
clearly showsthat the organizationof beneficiaries,both before and
after the enactmentof reform, is an indispensablecondition for its
success.
10. It should be recognizedthat landlessrecipients of land who
take up independent farming for the first time may need to be pro-
vided with their entire short-term and long-term credit requirements
and perhapssome consumption credit for three or four initial crop
seasons.There may also be a need for special training facilities,
researchactivitiesand field demonstrationsin suchcircumstances.
11. The abolition of tenancymay not be feasiblein manycountries
or regions where the demand for land by the landlessand small
farmers far exceedsthe availablesupply. In such cases,regulation of

12
tenancy might be a more efficient policy. Generally,fixed cash-rent
contractsaresuperiorto crop-sharingcontractsbecausethey encour-
age the use of inputs to the optimal level. But where crop sharing
cannot be eliminated becauseit provides risk insuranceto share-
croppers,it canbe mademore efficient andequitable if it is combined
with cost sharing.Suchcontractsshould be promoted with a system
of incentivesand deterrents.The incentivescan include the accrual
of legal rights in land and the availability of credit and other inputs
only if preferred types of tenancy contracts are implemented.
12. When the land-labor ratio becomesfavorable, the conversion
of tenants into owners of the land they cultivate, preferably against
very low compensationpayments,should be undertakenbecause,in
general, owner-operatedfarming is likely to be more efficient and
equitablethan tenantfarming.

World Bank'sPolicyGuidelines
1. The World Bankwill give priority in agriculturallendingto those
member countries that pursue broad-basedagricultural strategies
directedtoward the promotion of adequatenew employmentoppor-
tunities, with special attention to the needsof the poorest groups.
The Bank will support policies of land reform designed to further
theseobjectives.
2. The Bankwill makeit known that it standsreadyto finance spe-
cial projects and programsthat may be a necessaryconcomitant of
land reform, so long as the reforms and related programsare consis-
tent with the objectivesstatedin the previous paragraph.Thesepro-
gramswould include credit, technicalservicesand infrastructureproj-
ects designedto meet the specialneedsof land reform beneficiaries.
3. TheBankwill cooperatewith the FoodandAgriculture Organiza-
tion of the United Nations (FAQ),the United Nations Development
Programme(UNDP)and other organizationsto provide support and
assistanceto member governmentsseekinghelp with the specifica-
tion and design of land reform programswhere theseare in keeping
with the Bank'sobjectives.This support will include financial and
technical aid with cadastralsurveys,registration of land titles and
similarservices.
4. The Bankwill continue to explore, through its agricultural and
rural development projects, ways of providing for a distribution of
benefitsconsistentwith the goalsoutlined under (1) above,including
appropriate tenurial arrangementsand projects designedto servethe
needsof smallfarmersand settlers.

13
 5. The Bank will intensify its efforts through sector and country
economicwork to identify and draw attention to the need and oppor-
tunities for land reform with respect to existing tenurial situations
and their economiceffects.
6. The Bank will support and encourageresearchrelated to the
economicsof land reform in its broadestaspects,including its social
dimensions. It will continue its support for programsof economic
and technical researchdirected toward the specialneedsof the type
of small farmerlikely to emergefrom landreforms.
7.The Bank will undertake studies of the costs and benefits of
settlement projects, with particular attention to developing ap-
proacheswhich will lowerthe cost per family settled.
8. The Bank will not support projects where land rights are such
that a major shareof the benefits will accrueto high-income groups
unlessincreasesin output and improvementsin the balanceof pay-
ments are overriding considerations;in such cases,it will carefully
consider whether the fiscal arrangementsare appropriate to ensure
that a reasonableshare of the benefits accruesto the government.
9. In circumstanceswhere increasedproductivity can effectively
be achievedonly subsequentto land reform,the Bankwill not support
projectswhich do not include land reform.
10. Where land is held under someform of tenancy,the Bankwill
foster the adoption of tenancy conditions and sharecroppingar-
rangementsthat are equitable and conducive to the optimal use of
resources.
11. Where land is communally held without regulation of access,
the Bankwill encouragesubdivision,if sedentaryforms of agriculture
are possible,or pursue land usageand accessarrangementsthat are
compatiblewith the long-run productivity of the land and the welfare
of the residentpopulation.
12. The Bankwill pay particular attention to the consequencesof
the interaction of new technology and the prevailing institutional
structures,as reflected in the pattern of landownership,in order to
avoid adjustmentswhich will increasethe maldistributionof income
and causeeconomichardship.

14
 Chapter 1: CHARACTERISTICS
OF
LAND REFORM
Manand Land
Man's relationship to land, and patternsof landholding and land
use,are shapedby the interaction of a complex of forces-climatic,
economic, cultural, religious and political. In EasternAfrica, for
instance, physical conditions in the temperate areasare suited to
sedentaryagriculture, whereasthe more tropical and arid areasare
better suited to shifting cultivation or livestock herding.As a result,
different systemsof land managementand patternsof holdingshave
emerged in adjacent zones. Similarly, laws and customsgoverning
inheritancehavean effect on the distribution of land. Where land is
inherited by the oldest heir and not subdivided, the patternof hold-
ings is lessfragmentedthan in societieswhere the customis to divide
holdings equally among all heirs. In addition, many socioeconomic
factorsaffectcustomsof usufruct,traditions of crop sharingand other
arrangementssurroundingland usein varying situations.
The political ideologiesof governmentsalso havea bearingon the
relationshipbetween people and the land.The right of the individual
to own, sell and accumulateprivateproperty-including land-is one
of the cornerstonesof the market economy.While this right might be
constrainedin the public interest, land can in generalbe exploited,
held and traded by individuals for private gain. Under some other
ideologies, individuals do not have the opportunity to acquire and
accumulateland; the right to own land may be vested solely in the
stateor in semipublicinstitutions, and it is the statewhich organizes
and controls the land accordingto its own criteria. To the extent that
the statecontrols the land,the allocativeprocessmayserveany num-
ber of ideological ends.Somegovernmentshaveused control over
land to implementpoliciesof geographicalseparationof racialgroups.
The People'sRepublicof China,on the other hand, haschangedrights
to land and the organizationof work severaltimes over the past 25
yearsas part of a drive to eliminate rural inequality.
The level of economicdevelopmentof a country hasa strong influ-
enceon attitudestoward land.The more industrializeda country, the
smaller the proportion of the population in agriculture and the less
significantthe role of land in the economy.In countries with mobile
populations which have ample opportunities for employment,land
is often seenmerelyasone factor of production in a highly developed
commercial agriculture. However, in less developed countries with
large rural populations,limited alternativeopportunitiesand increas-

15
ing pressureon the land, accessto land may provide at leasta sub-
sistenceincome. In thesecircumstances,producerssee landas more
than a factor of production; it may well provide the margin between
destitution and subsistence.
The establishedpattern of landownershipis basicto both the social
organization and institutional structures in rural areas. The social
hierarchy in most agrariansocietiesreflectsthe kinds of accessthat
different groups have to land, while individual status within these
groupsdependson the amountand quality of land commanded.The
institutional structureswhich formalize the various meansof control
and the relationshipbetweencategoriesof land users,alsodetermine
the accessibilityof external institutions and servicesto the various
groups.

Contextof LandReform
The many complex factors that influence the patterns of land-
ownership and land use in different regions of the world may be
summarizedas: (1)the political systemand situation; (2)the structure
of the economy; (3) the social system; (4) the legal system;(5) the
demographic situation; (6) the agricultural system; and (7) the
national resourcebase.When these interacting elements are taken
into account, it is possibleto delineate six main categoriesof land
tenure and land use.These are characterizedas follows:
1. FeudalAsian Type
High property concentration.
Great socialinequality.
Greateconomicinequality.
Low land productivity.
Low labor productivity.
Low level of technology.
Mainly operatedby sharecroppers.
High labor intensity.
Low capital intensity.
Productionmainlyfor subsistence.
Landvery scarce.
Institutional structurecentralized.
2. FeudalLatinAmericanType
High property concentration.
Greatsocial inequality.
Great economicinequality.

16
Low land productivity.
Low labor productivity.
Low level of technology.
Labor provided by squatters,neighboring smallholdersand
migrantworkers.
Capital-extensive.
Labor-extensive.
Operatedby owner or managerplus hired labor, serfs
or sharecroppers.
Productionfor subsistenceand export.
Institutionalstructurehighly centralized.
3. Traditional Communal Type
Low property concentration-sovereign rightsvested in community.
Decentralizedcultivation-usufruct rights for membersof group.
Moderateor high socioeconomicequality.
Low labor productivity.
Low land productivity.
Low levelof technology.
Medium labor intensity.
Low capitalintensity.
Productionfor subsistence.
Supportingservicestructureunderdeveloped.
4. Market Economy Type
Medium property concentration.
Decentralizedcultivation.
Medium socioeconomicinequality.
High landproductivity.
High labor productivity.
High level of technology.
Capital-intensive.
Labor-extensive.
Market production oriented.
Institutionsand servicesdispersed.
5. Socialist Type
Propertyright vestedin the stateor a group.
Centralizedor decentralizedcultivation.
Low,medium or highsocioeconomicequality.
Low,medium or highland productivity.
Low, mediumor high labor productivity.
Medium level of technology.

17
Productionfor marketor subsistence.
Supportingsystemscentralized.
6. Plantation Ranch Type
High property concentration-owned by state or foreigners.
Great socialinequality.
Great income inequality.
High landproductivity.
Low or mediumlabor productivity.
Medium or high level of technology.
Operatedby managerpluswage labor.
Productionmainlyfor export.

In a traditional context, extremes in the pattern of land control

are exemplified,on the one hand, by the feudalistic landlord-tenant
systemfound in someAsianand LatinAmericancountriesand, on the
other, by the communal landownership pattern of certain tribal
groups in Africa. In the landlord-tenant system, landownership is
vested in an elite minority with the majority having accessthrough
tenancyarrangementsof various kinds.The ownership of property is
generally highly concentrated, more so than the pattern of land-
holdings. However, since holdings (the only category for which the
Bank has data) involve leaseholdunits for which rent is paid on a
share basis,the distribution of income is also highly skewed (see
Annex 1, Tables1:6 and 1:8). In the communalsystem,by contrast,
land is common property and accessto it is relatively unrestricted.
Whereasin the feudalistic systemthe distribution of landownership
and benefits are highly skewed and classdifferentiation is marked,
the communalsystemhas relativelyegalitarianland accessand class
differentiation is lessmarked.
Both systemsare relatively stable under favorable conditions, but
face difficulties as the man-land ratio declines through population
growth, unlessthere are offsetting changesin technology. In the
landlord-tenant system, land pressuresare reflected in a growing
army of landless people and widening income differentials (see
Annex 1, Table1:11).The communalsystemmanifeststhe samepres-
suresby compressedfallow periods and declining soil fertility, over-
grazing and increasederosion, accompanied by extensive poverty
and vulnerability to seasonaleffects.
The two systemsdiffer in their ability to respond to changing
external conditions and especiallyto new technology. The landlord
elite, by virtue of its privileged position and power, can, and often
does, becomeeducatedand innovate both through experimentation

18
and the adoption of externalideas.(in doing so,however, its primary
concern may be to promote its own narrow interests in terms of
wealth and power, for instance,by displacingtenantsthrough mech-
anization.)Thecommunalsystemgenerallylackssuchan institutional
mechanismand tendsto be both static in itstechnologyand relatively
insular, but such communitiesseldom manageto remain completely
isolatedfrom externalinfluences.
In a modern context, the extremesin patternsof land control are
seen respectivelyin the private ownershipof land, which is a funda-
mental aspectof the market economy and common in mostWestern
countries, and the state or collective ownership characteristicof
socialist countries. Under private ownership, land is held by indi-
viduals and, while usually subject to special restrictions, can be
bought or sold like any other commodity. Suchholdingsare typically
operated as family units with little hired labor. However,a rangeof
subtypesexistswithin this categorywhich reflectsa gradation in size
from the predominantly subsistencesmallholdings of many devel-
oping countries to the broad acresof North America and Australia.
Although similar in legal and institutional respects,thesediffer sig-
nificantly in their technologyand input mix aswell as in the degreeof
market orientation.
In the socialistsystem,on the other hand, little or no provision is
made for individuals to acquire or accumulateland, this right being
vested in the state,with control determined in accordancewith the
objectives of the state. But some variations remain within many
socialist systems,often providing for the existenceof private small-
holdings in parallel with larger social units. A special type found in
a modern context is one which includes the plantations and large
ranchesthat often operate in developingcountriesas well as in some
developedcountries.Theseform, in somerespects,a specialcategory
of the market economy type, but the tendency toward a corporate
legal structure and dependence on hired labor differentiate them
from privatelyowned family farms.
While private ownershiphasgenerallybeen compatiblewith tech-
nological progressand the economicadjustmentof agriculture,it has
often created inequities as people have been compelled to give up
rural pursuitsor havebeen squeezedinto land-scarcerural enclaves.
Generally,private control has been most satisfactorywhere popula-
tion pressurecould be offset by colonizing virgin land or moving
people out of the rural sector. It hasbeen most unsatisfactorywhere
ownership patternshave become skewedbecauseof the growth of
large farms,combinedwith limited opportunities for peopleto move
out of agriculture, and the subsequent emergenceof economic

19
dualism. Stateor communal control has led to fewer interpersonal
inequities, although in most casesnot without some broader eco-
nomic inefficiencies.
Land reform raisesissuesof equity in the context of both the tradi-
tional landlord-tenant relationship and the modern skewed owner-
ship pattern. In both these contexts, it is often a highly political
concern, especiallyin the traditional feudalistic and communalsys-
tems. In many situations, the prevailing tenure conditions are the
major impediment to development. For example, a high level of
fragmentation can make canal irrigation virtually impossible and
seriouslyimpede mechanizedoperationseven when on a very small
scale.In other cases,the contractualsharearrangementis such that
neither landlord nor tenant are able to introduce new technology
because,on the one hand, the landlord cannot capturea profitable
share of the return on his investment,and on the other, the tenant
cannot find the capital for investmentor lacksthe securityof tenure
that would guaranteea return from it. Further, in some situations,
the social environment is characterizedby inequity and oppression
to the extent that it destroyshuman motivation to improve produc-
tivity or to resolve any problem within existing structures. In such
circumstances,land reform may become a prerequisiteof develop-
ment. But, whether primarily an equity or a production concern, it is
clear that land reform will involve different changesin different types
of situations.

Dimensions of LandReform
Land reform is thus concerned with the interrelated aspectsof
productivity and equity of land use.It is frequently pursuedas a goal
in itself, but in a development context is usually seen as a part of
agrarian reform or of rural development programs. Land reform
differs from political, administrative,fiscal or monetary reforms in
that it normally relatesto one sector and involveschangesin control -
of a tangible assetthat not only is fixed in supply but also provides
the basicfactor on which most of the people in developing countries
dependfor their livelihood.
Landreform caninvolve varyingdegreesof change,including some
or all of the following:
1. Redistributionof public or private land in order to changethe
patterns of land distribution and size of holdings. Usually,this in-
volves an increasein the number of small- or medium-sizedfarms
and a reduction in the number of large holdings. Alternatively, all
land can be nationalizedand regrouped into state-ownedholdings,
all of which might be large.

20
 2. Consolidation of individual holdings, thereby reorganizing the
physicalpattern of control. Fragmentedholdings can be regrouped
into contiguous blocks of land. This can be done with or without
changing the distribution of landownershipin terms of acreageor
valuebelongingto eachindividual.
3. Changesin landownershipand tenurial rights,with or without
physical redistribution of land. Redistributedland can be allocated
to new owners or to farmersworking on the land.Alternatively,land
need not be redistributedbut tenantsor workerscanbe madeowners
of the land they work. In that case,the resultis generallya redistribu-
tion of income away from the former owners of the land to the new
owners. The new owners may farm cooperatively or as individuals.
4. Changesin conditions of tenure without changing ownership
or redistributingland. The rightsof thoseworking on the land can be
safeguardedby law without a changein ownership.Changesin con-
ditions of tenure would include providing security of tenure, intro-
ducing equitable crop-sharing arrangements, cooperative land
management,and so forth. These changeswould also include the
conversion from customaryto legal rights to land.
Structural Change
In the main, land reform is seenas a meansof bringingabout struc-
tural changes in the agricultural sector, thereby altering the size
distribution of holdings or the distribution of income. By definition,
therefore, pilot projects cannot be consideredto be land reform for
they operate within an existing structural framework, even though
they might be useful in identifying problems of management,or the
economicsof various "models," or arrangementsthat might be part
of a subsequentreform.Similarly,landsettlementon the frontier does
not usually constitute land reform, although land settlement might
be a meansof bringing unusedlandinto production. Landsettlement,
by itself, may or may not have an impact on the structure of land-
holdings in a country, dependingon the manner in which the settlers
are selectedand the size distribution of the new holdings.The kind
of structural changeinvolved dependson the prevailingtenure type
and the proposedalternative.As reflected in the country experiences
summarizedin Annex2, most changesinvolve a shift from traditional
to modern types.Thus the Republicof China,the Republicof Korea
and Japanmoved from a "feudal Asian" to a "market modern small-
holding" type; India and Iran moved from a "feudal Asian" toward
a "market modern" type, with some traditional farms retained and
some "plantation ranch" type variations in certain areas.Kenyaand
Morocco redistributed the large-scale,alien-owned "market econ-
omy" type holdings of their colonial eras,some going to smallhold-

21
ings of the "market economy" type and someto "plantation ranch"
type units. Mexico and Perumoved from a "feudal Latin American"
type to a "market modern mixed large and smallholding" type, and
a mixed "market modern" and "socialist" type structure,respectively.
Thesechangesin tenure systemswere in all casesaccompaniedby
changesin relatedorganizationsand services.
FiscalMeasures
Land taxesand preemptive taxeson income earnedfrom land are
often cited as instruments that will obtain the sameends as land
reform. An effective land tax mayhavean impact on land usebut its
main purpose is usually to encouragemore intensive production by
making it costly either to leaveproductive land idle or to useit below
its productive capacity. On the other hand, such taxesmay provide
a disincentiveto investmentwith the potential of increasingproduc-
tivity or bringing new land into production. In any event, the useof
a fiscal instrument, such as a land tax, will not lead to structural
changesin agriculture-at least not in the short run. A more likely
fiscal instrumentto encouragestructuralchangeis a graduatedestate
tax which would force estatesto disposeof land to meet their finan-
cial obligations.But this is likely to bring about structuralchangeonly
over a long period of time. While landtaxesand estatetaxesare often
considered significant elements in fiscal policy intended to redis-
tribute income, they cannot ensure the same degree of structural
reform as can land reform and have,in general,been quite ineffec-
tive. In situationswhere fiscal measures-whether of a redistributive
kind or a typewhich providesa returnto the stateon its investment-
are found to be ineffective, land reform may be the only alternative
option if economicdevelopmentisto be pursued.
AgrarianReform
Agrarian reform is a much more comprehensiveconceptthan land
reform, since it involvesmodification of a wide range of conditions
that affect the agriculturalsector.Thesemodificationsmight include
changingprice policiesso as to turn the terms of trade in favor of the
agricultural sector; increasingallocations to the agricultural sector
in order to expandresearch,extension,training and storagefacilities;
making physicalsupplies,suchas fertilizers, availableand increasing
credit for their purchase;or providing infrastructureto facilitate agri-
cultural production. Agrarian reform may or may not include land
reform; in some instances,there may be no need for land reform
since land is alreadyevenlydistributed. In other cases,it maynot be
politically feasible to have land reform-although it might be both

22
politically and economically feasible to raise output through the
measuresinvolved in agrarianreform. The point is that land reform
may be a necessarycondition for agrarianreform, but it is seldom a
sufficient condition for increasingagricultural output, since land is
only one factor of production.
RuralDevelopment
Broader still is the concept of rural development,becauseit em-
bracesall dimensionsof the rural sector (agricultural and nonagri-
cultural) and is more concernedwith the welfare of rural people than
with agricultural output or productivity as an end in itself. Since it
hassignificant equity implications, land reform may be a necessary
concomitant of successfulrural development,dependingon the pre-
vailing pattern of land control. Where the ownershipof land directly
affects the nature of local institutions and the participation in them
by the majority of rural people, land reform maybe essential.How-
ever, in termsof implementation,in somesituationsestablishinglocal
institutions and smallholder servicesmay be a prerequisite of land
reform rather than vice versa.Where the existingservicesystemsand
administrativestructureis gearedto working with large-scalefarmers,
land reform without concurrent rural development activity might
causehardshipand economic losseswhich would outstrip the equity
gains associatedwith land redistribution. Tenancy reform, on the
other hand, insofar as it stabilizesthe existing relationship between
landownersand renters,may be a useful precursorof rural develop-
ment programs.
PoliticalDimensions
Substantialreform of the structureof holdingsand the distribution
of income from the land cannotbe achievedwithout political action.
For instance,where semifeudalconditions prevail, patterns of land
rights and tenurial conditions havebeenestablishedby tradition, and
thesecannot be changedthrough market operations,as there is vir-
tually no organized market for land. Elsewhere,large landholders
have accumulated capital and expanded landholdings acquired
through the market; in most market-oriented economies with a
skeweddistribution of land, the tendency is for the skeweddistribu-
tion to worsen. Whatever the prevailing situation, it can seldom be
changedwithout actionsthat emanatefrom outside the market.Since
theseactionsare basedon policiesdeliberately intended to alter the
distribution of land and change tenure, the implementation of the
policies dependson the political will of the policy makersand the
ability of the administratorsto executethis will.

23
 The concentration of control over land provides the base for
powerful elementsin manynonindustrializedsocieties.Where groups
derive authority from their land, a meaningfulland reform program
will inevitablydestroy or limit the power baseof thesegroups.Land
reform can changethe political balance and the power structure in
a country. Reforms have stripped large landholders,whether they
were military, religiousor private,of their power. It is not surprising,
then, that land reform is often a central issuein political debatesand
that thesedebatesare often couched in termsof redistributingpoliti-
cal power as well aswealth. The political implications of land reform
must be taken into account; ambitious programs of land reform
will seldom be implemented unlessshiftsare made in political senti-
ment and power. Many countries have legislatedfor land reform but
relatively few have achieved it-and these only with a change in
government.
Frequently,the implementation of massivereform legislation has
dependedon the effectiveorganizationof the beneficiaries.In Japan,
the Republicof China and Venezuela-to name three countries-
suitableorganizationswere establishedto ensurethat landwas indeed
transferred.In other countries,suchas India and Pakistan,the official
bureaucracywas the only implementation agencycontemplated by
the reformers. Becauseof the community of interests between the
bureaucratsand the landowners,and the absenceof organizedpres-
sure from the beneficiaries,the massivelegislationhasproduced no
significantreform. Experiencein much of Asiaand LatinAmericasug-
geststhat effective popular participation of rural people may be a
critical condition of successfulland reform.
Implications for Social Justice
The imbalancebetween the distribution of control over the land
and the numbers dependent on it has historically led to increasing
pressuresfor change.While the focus on land reform is related to
economicdevelopment,the concept of an overriding social function
of land justifying the imposition of limitations on private rights ap-
pears to be gaining the support of many groups, including the
CatholicChurch. Formerlyone of the largestlandholdersin the world,
the Church in Europeas well as in LatinAmerica hasincreasinglyput
its weight behind this new concept, both in precept and in practice.
The Church's new philosophy regarding the relationship between
man and land declared that "private property doesnot constitutefor
anyone an absolute and unconditional right." And the immediate
extensionof this postulateto the world's agrarianproblem is that "if
certainlandedestatesimpedethe generalprosperitybecausetheyare

24
extensive,unusedor poorly used,or becausethey bring hardship to
peoplesor aredetrimentalto the interestsof the country,the common
good sometimesdemandstheir expropriation."
A further facet of land reform that warrantsconsideration in this
respectis the potential of a new societalstructurefollowing a reform.
Mexico, and more recently Bolivia and Egypt,had semifeudalsocie-
ties similar to many which still prevail in other parts of the world. In
these societies,large numbersof tenants and laborerswere tied to
the land and were held in forms of human bondage;this arosefrom
custom, tradition or sheer indebtednessto landlords. The reforms
which havetaken placein thesecountrieshavechangedthe situation.
The reform in Mexico broke a systemthat denied many people any
range of choice in the pursuit of a livelihood. If the experienceof
Mexico-which hashad the longest period of reform-is any indica-
tion of the long-run outlook, the reforms haveled to an increasein
socialmobility.
Land reform is a complex subject.The issuesinvolved are diffuse
and appropriate reform measuresvary according to the situation.
Land reform is in practice predominantly a question of equity and,
therefore,one that is often highly political. Nevertheless,it hassignifi-
cant implications for economic development,and thesein turn are
relevantconcerns in the formulation of the World Bank'spolicy.

Chapter 2: LAND REFORMAND

ECONOMIC DEVELOPMENT
Economicdevelopmenthasthree basicobjectives:rapid economic
growth, full employmentand distributive justice. Somepolicies and
related investments,such as those affecting power plants or large-
scale industry, are primarily growth oriented; others, such as those
for rural works, are employment oriented; still others, such as those
related to land reform, are essentiallyequity oriented. Eachset of
policies and investmentsaimed toward one objective has important
repercussionswith regardto the other two objectives,and thesemust
be taken into accountwhen weighing the potential impact of particu-
lar policieson economic development.Forthis reason,it is important
to determine to what extent land reform might be costly in terms of
growth and employment.
Many problems arise in assessingthe costsand benefits of land
reform. Theseinclude the definition of an acceptabletime frame for
measuringthe effects of the related structural changein the agricul-

25
 Table 1

Productivity, Employment and the Distribution

of Land, in Selected Countries

FarmGDP FarmGDP Sizeof Gini's

per per Employment average Indexof
Data hectare worker per holding Land
Country year (US$) (US$) hectare (hectares) Concentration

Europe
Greece 1961 424 848 0.50 3.18 0.597
Spain 1962 90 980 0.09 14.85 0.832

Central America
CostaRica 1963 83 951 0.09 40.70 -
DominicanRepublic 1971 129 463 0.28 8.64 -
El Salvador 1961 186 489 0.38 6.95 -
Guatemala 1964 144 492 0.29 8.17 -
Mexico 1960 22 569 0.04 123.90 -
Nicaragua 1963 55 580 0.09 37.34 -

SouthAmerica
Argentina 1970 18 1,903 0.01 270.10 0.873
Brazil 1960 14 285 0,05 79.25 0.845
Chile 1965 18 692 0.03 118.50 -
Colombia 1960 67 663 0.10 22.60 0.865
Paraguay 1961 11 479 0.02 108.70 -
Peru 1961 50 477 0.10 20.37 0.947
Uruguay 1966 14 1,333 0.01 208.80 0.833
Venezuela 1961 31 925 0.03 81.24 0.936

Asia
China,Republicof 1960-61 841 410 2.05 1.27 0.474
India 1960 172 141 1.22 6.52 ' 0.607
Indonesia 1963 323 149 2.17 1.05 -
Iran 1960 187 581 0.32 6.05 0.624
Korea,Republicof 1970 1,085 377 2.88 0.85 -
Japan 1960 1,720 1,188 1,45 1.18 0.473
Nepal 1961-62 352 138 2.54 1.23 -
Pakistan 1960 240 249 0.96 2.35 0.607
Philippines 1960 250 200 1.25 3.59 0.580
Sri Lanka 1962 376 337 1.12 1.61 -
Thailand 1963 166 137 1.21 3.47 -
Turkey 1963 155 243 0.64 5.03 0.611
Viet-Nam,Republicof 1960 355 127 2.79 1.33 -

Africa
Botswana 1969-70 168 142 1.18 4.75 -
Egypt,ArabRepublicof
1960-61 681 360 1.89 1.59 -
Kenya 1969 183 140 1.31 4.20 -
MalagasyRepublic 1961-62 293 88 3.32 1.04 -
Mali 1960 98 48 2.06 4.35 -
Morocco 1961 144 295 0.49 4.62 -
Senegal 1960 209 174 1.20 - 3.62 -
Togo 1961-62 189 180 1.05 2.62 -
Tunisia 1961-62 42 341 0.12 15.41 -
Uganda 1963-64 167 198 0.84 3.29 -
Zambia 1960 68 101 0.67 - -

Sources:Columnsland 3arebased on FAO,PreductionYearbook1971.pp. 10-11, 21-23, and column4onUN, Monthly

Bulletin XXVI, No.4, April 1972,and XXVII, No.11, November1973.Forcurrencyexchangerates,see
of Statistics,
ibid, and I MF, lnternationalfinancialStatistics.XXVI, No. 3, August 1973.GrossDomesticProduct(GDP)in agriculture
shownhere includes,unlessotherwise indicated,agriculture,hunting, forestry,and fishing.

26
ture sector.The availableevidencesuggeststhat a well-designedland
reform program need not entail unacceptablecostsin termsof other
objectives; its contribution to output and employment-as well asto
equity-depends on the speedand effectivenessof the reform and
complementary investments.However, the effects of land reform
canbest be examinedby focusingon particular measures,suchasthe
effectsof farm size on productivity,equity and employmentaswell as
on savingsand market surplus.These measuresare interrelatedbut,
for analytical convenience,are treated separatelyhere.

Implicationsfor Productivity
The effects of land reform on productivity might best be isolated
by comparing productivity in a given area before and after reform.
Unfortunately, this is not possible as there is no situation where
changehas occurred in only one variable-size of farm-over time.
The nearestalternativeis the comparisonover a definedperiod of the
productivity of groups of different-sizedfarms in a given area.The
ideal measurefor comparisonwould take into account the contribu-
tions of all factorsof production and so measuretotal factor produc-
tivity. Since data are not availableto derive this measure,changesin
yields per hectare are considered to be the most appropriate
substitute.
Several comparative multicountry analyseshave been made of
the effect of differencesin distribution of size of holdings on yields.
One 13-country study undertakenby the FAO analyzedthe relation-
ship among size of holding, concentrationof land and productivity.
A similar study of 40 countries was undertaken by the Bank (see
Table1). Both studiesindicatedthat a smalleraveragesizeof holdings
and a lower concentrationof landownershipwere associatedwith an
increasein output per hectare.
Similar findings can be cited from cross-sectionstudiesin a num-
ber of individual countries.In Sri Lanka,for example,in 1966-67,the
yield of paddy averaged36 to 37 bushelsper acre on farms of up to
one acreand 33 to 34 bushelson largerholdings.In central Thailand,
yieldswere reportedto decline from 306kilogramsper raion holdings
of two to six acres,to 194 kilogramsper rai on holdings of 140 acres
or more (1 rai equals0.4 acre).Small farms in the Philippines-that
is, farms of less than two hectares-produced 2.9 tons of paddy per
hectare,while farms of more than four hectaresproduced 2.2 tons
per hectare.In a systematicanalysisof the differencesbetween large
"multifamily" farms and small "subfamily" farms in Argentina,Brazil,
Chile, Colombia, Ecuadorand Guatemala,output per hectare was

27
 Table 2
AgriculturalOutputper Hectareand per Worker,
by FarmSize, in LatinAmerica
1 2 3
Smallest Largest Ratio ot
subfamily multifamily col. I to
Country Year farms farms col. 2

National
monetary
unitperagricultural
hectare
Argentina 1960 2,492 304 8.20
Brazil 1950 1,498 170 8.80
Chile 1955 334 41 8.20
Colombia 1960 1,198 84 14.30
Ecuador 1954 1,862 660 2.80
Guatemala 1950 63 16 3.90
National
monetary
unitperworker
Argentina 1960 40 192 0.21
Brazil 1950 1,197 8,237 0.14
Chile 1955 268 1,171 0.23
Colombia 1960 972 9,673 0.10
Guatemala 1950 74 523 0.14

Source: Barraclough and Collarte. Agrarian Structure in Latin America, a resume of the CIDA Land Tenure Studies of
Argentina, Brazil, Chile, Colombia, Ecuador, Guatemala, Peru. xxvi, 351 p. Studies in the Economic and Social
Development of Latin America. Lexington, Massachusetts: Lexington Books, 1973.

found to be three to 14 times greater,on the average,on the small

farmsthan on the largefarms (seeTable2).
There is other evidence to support these findings, including the
results of Bank-sponsoredanalysisin Mexico, as well as studieson
Japanand the Republicof China. However,there is no claim that all
conditions were identical; the studiessimply indicatethat yieldswere
higher on smallfarmsthan on largefarms.
The important implication is that reductionsin either the size of
holdingsor land concentrationneed not be associatedwith a reduc-
tion in output per hectare.On the contrary, it appearsthat under
controlled circumstancesoutput per hectare is likely to be higher.
There are two associatedreasonsfor this assumption.Firstly, there
are limited economies of scale in most agricultural production.-
Secondly,small-scaleproducerstend to maximizeoutput by applying
labor intensively,while large-scaleoperatorstend to maximizeprofits
by using hired labor only until incrementalproduction coversincre-
mentalcosts.This is usuallyshort of the output per hectarethat would
be produced if the goal were maximization of output.
In broad terms, land reform can be consonantwith development
from a point of view concernedpurely with productivity,with output
per hectareas the relevantcriterion. Output per worker, however,is
likely to decreasefor the simple reasonthat, as pointed out below,

28
smaller farmswould employ more labor per hectare.In other words,
the larger income would be sharedby an evenlargernumberof fami-
lies. This decline in labor productivity only reflectsthe employment
and equity benefitsof land reform: the samelandwould supply more
people and the income generatedwould be more widely shared.

LandReformand Employment
Evidenceexists that the use of labor per hectare is greater on
smaller holdings than on larger ones.The cross-sectionalanalysisof
the 13 countries previously mentioned shows that manpower per
hectareof agricultural land is significantlycorrelatedwith the size of
the holding-the smaller the holding, the greaterthe input of man-
power. This cross-sectionalevidence of the higher productivity of
small farms indicates their long-run equilibrium potential. But the
realization of this potential is contingent on the supply of nonland
inputs being increasedas soon as farm size is decreased.
A limited number of studies in Asia and Latin America have also
confirmed these findings. In the Ferozepurdistrict in Punjab (India),
for example, in 1968, labor absorption varied between 33 and 39
man-days per acre on holdings of less than 30 acres. On larger
holdings, it ranged from 20 to 23 man-daysper acre. In Colombia,
man-yearsper hectare declined steadily from 2.7 on small holdings
(lessthan 0.5 hectare)to 0.17 on large farms (500to 1,000 hectares)
in 1960. In other Latin American countries (Argentina,Brazil, Chile
and Guatemala),the number of workers per hectareof agricultural
land on the smallestfarms (subfamilyunits) hasbeen estimatedto be
30 to 60 times greaterthan on the largest(multifamily) farms.
More intensive labor use is the main reasonwhy small farms are
ableto producemore per unit of landthan the largerfarms.But inputs
other than labor arealso likely to be applied more intensivelyon small
farms, unlessaccessto these inputs is blocked by institutional ar-
rangements.Unfortunately, the relationship between these other
inputs and farm size cannotbe studied in manydeveloping countries
JFor want of data. It is interestingto note, however, that in the cross-
section of developed countries,in 1961, fertilizer consumption and
grossfixed capital formation per unit of landwere relativelyhigher in
countrieswith smalleraverageholdings.
In developingcountries,too, small farms undoubtedly need much
more nonlabor input in order to raiseproductivity.The mereredistri-
bution of land and increasein employment may not suffice to raise
output substantially.Therefore,the organization of an effective ex-
tension-cum-inputsupply systemfor small farmersmust accompany

29
land reform. Where there is such a system-as in Japan,the Republic
of Koreaand the Republicof China-the absorptivecapacityof agri-
culture tends to be high even though holdingsare small; at the same
time, output per hectareis high. Smallholdingscanyield high returns
to labor provided output per hectare is high-a condition that can
only be fulfilled by the application of high-yielding, labor-intensive
technologies.

LandReformand Equity
The more radical the land reform and the more important the
share of agricultural land in relation to total tangible wealth, the
larger will be the equity effect of the reform program. In the rural
areas,agricultural land accountsfor such a large proportion of total
wealth that it is usuallythe single mostsignificantdeterminantof the
distribution of both income and power. Evidenceof this can be seen
in many LatinAmericanand Middle Easterncountrieswhere the large
landownersoften dominate both commerceand government.There,
land reform could havea major equity impact. However,where much
of the wealth existsin the form of financialassets,real estateand other
investmentsapart from farmiand,and commodity stocksin the hands
of traders,the redistribution of farmland alone may not improve the
distribution of total wealth substantially. Landownersmay easily
changethe composition of their assetson the eve of land reform if
agricultural land alone is the target of redistributive zeal.
If rural and urban areasare consideredtogether, the limitations of
redistributingfarmlandaloneappearevenmore serious.Thedistribu-
tion of real estate,financial assetsand commodity stocksin the urban
areas is even more skewed than the distribution of farmland in the
rural areas.If, therefore, urban property reform or highly progressive
taxation on urban wealth does not accompanyland reform in coun-
tries with a substantialand prosperousindustrial-commercialurban
sector, land reform alone is not sufficient. By itself, it not only may
not decreasethe inequity of the distribution of total wealth in the
country as a whole. It may even increasethe inequity-in particular,
the inequity between the town and the village-since it will freeze
the maximum permissibleownership of the main rural asset,without
freezing the maximum permissibleownership of urban assets.
Evenwith this broader focus, the equity effect of land reform will
be significant only if: (1) the effective ceiling is low; (2) the benefi-
ciariesbelong to the poorer groups; (3) the extensionand (nonland)
input distribution systemfavorsthe beneficiaries;and (4) owned and
self-operatedland as well as leasedland is redistributed.

30
The PopulationFactor
Opportunities for the redistribution of land depend to a great ex-
tent on the existingpattern of distribution of holdingsand population
density. As will be shown later, there are some countries, notably
in the Americas,where land distribution is skewedand population is
not dense.In suchcountries,there areample opportunitiesfor redis-
tributing land so that inequalities can be diminished and the recipi-
ents of the land can generate an acceptable minimum income. In
other areas,however, the pressureof population is such that there
is not enough land to meet the minimum requirementsof all claim-
ants.The densityof the farm sector is so high in some countries in
Asiathat, evenif holdingsabovea certainsize werecompletely elimi-
nated, not enoughland would be availableeither to raisethe acreage
of the minifarms to a tolerable minimum or provide for the landless.
In India, even if the maximum holding was 20 acres,the available
land (43 million acres)would be barelysufficient to bring up the size
of miniholdings to a minimum of five acres,and no land would be
availablefor the landless(20-25million households).In Bangladesh,
a low 10-acreceiling would not sufficeevento bring all miniholdings
up to a minimum two-acre size. The millions of landless families
could not be provided for at the sametime. In Sri Lanka,too, even
with a low ratio between the ceiling and the floor holding (5 to 1),
therewould be enoughlandonlyto givetwo acrestoeachminifarmer.
In Haiti, only 1.5 hectaresis availablefor the averagerural family of
five. The solution to rural povertyclearlycannot be found exclusively
in the agriculture sector. In thesesituations,it might be wise to give
land only to the minifarmersand to attackthe poverty problem of the
landlessby meansof a massiverural works program. (Settlementof
the landlesson new land, where available,and their migration to
urban areas,when possible,are the other obvious alternatives.)

.Effectson MarketedSurplusandSavings
The redistribution of land can have a pronouncedimpact both on
the availability of a marketablesurplus and on aggregatesavingsin
rthe agriculturalsector.Although the total effect of the redistribution
processwill dependto a largeextent on the costsof increasedoutput
after the redistribution,the changein the sizedistribution of holdings
will shift the distribution of the sourceof the marketablesurplusand
savings.As the marketedsurplus generatesagricultural incomesand
so potential cashsavings,it determinesthe size of the rural market
for domesticallyproduced industrial products.The marketedsurplus
also representsthe supply of agricultural products, mostly food, for

31
the urban population. Thus, a fall in the surplus could necessitate
imports and put an added strain on the balance of payments.But
increasingthe marketedsurplus will not necessarilyincreasesavings.
Where it does,the savingsneed not be monetized,but may take the
form of increasedon-farm investment in such items as improved
housing,wells and accessroads.

MarketedSurplus
A reduction in land concentrationthrough land reform could lead
to a fall in the marketedsurplus-at leastin the short run. Smallfarm
householdstend to consumea largerproportion of their smalloutput
than do householdswhich havea large enoughacreageto produce
in excessof domestic requirements.Thus, the ratio of marketedsur-
plus to production falls asfarm size decreases.Datafrom India show,
for example, that small farms (2.5 acresor less)sell only 24.5% of
their output, whereaslarge farms (50acresor more) sell 65.4%. But
thesefarm groups produce only 9.5% eachof the national output. If
output remainedthe samebut, hypothetically,farms abovea certain
size were eliminated and their land transferredto the small class,the
surplus-output ratio would probably decline. The rate of decline,
however, might not be very great given that the largest and the
smallestfarm-sizegroups account for only small proportions of the
total output.
The surplus-outputratios of different farm-sizegroups, however,
and their sharesof total output and salescan differ widely across
countries and regions. Sixty-one percent of the maize farmers in
Puebla(Mexico),for example,sell no maize at all; and another 16%
sell 25% or lessof their output. In Chile, on the other hand, a typical
sharecroppersells as much as 43% of his output. In Mexico, 6.6%
of the marketed surplus comes from 70.7% of the farmers; and
55.4% comes from only 1.7%. In India, 48% of the farms (lessthan
2.5 acres)contribute only 6% of sales,1% (more than 50 acres)con-'
tribute 16%, and 51% (with 2.5 to 50 acres) contribute the bulk
(78%)of the total surplus.
These differenceswould determine how much the surplus ratio
would fall after land reform; but there can be no doubt that it would
fall, with adverseeffects on the economy.However, this decline in
the market surplus ratio need not result in a decline in total surplus,
provided that there is a compensatoryincreasein total output. Since
per acreyields on smallfarms canbe higherthan on largefarms,there
may be a sufficient increasein output if, after reform, the necessary
conditions are fulfilled whereby small farmscan realizetheir full pro-

32
duction potential. In addition, from the welfare point of view, a de-
cline in the market surplus ratio has a direct distributive dimension
which should be offsetagainstthe decline.As the surplus-outputratio
falls, the subsistenceconsumption of small farmers increases-the
extra consumption in kind representinga direct increasein their in-
comes (nutrition). Insofar as the productivity of small farmers was
previouslyconstrainedby inadequatenutrition, there should also be
a positiveeffect on productivity.

Savings
In consideringthe productivity effect of land reform, it is necessary
to examinethe implications of a changein farm-sizestructureon the
aggregatesavingsrate of the farm sectoras a whole, sincethe savings
rate representsthe contribution of the sector to the long-run growth
of both its own productive capacityand that of the rest of the econ-
omy. Although the evidence on savingsrates of different classesof
farm householdsin developing countries is scant,it can be expected
that the behavior of the savingsrate will be similar to that of the
marketed surplus.At the lowest end of the farm-sizescale,the sub-
sistencefarmerscan be expectedto be net "dissavers"(for instance,
by running down the existing soil fertility). As farm size increases,
the savingsrate can be expected to become positive and increase
along with it (althoughlarge farmerscanbe "dissavers"too, by using
capital for consumption). A recent study in the state of Haryana
(India) tended to confirm this: the savingsratio was found to be
-0.24% for small farmers,8.5% for medium farmersand 16.3% for
large farmers.In a further study in Orissa(India), there was no direct
measureof the savingsmade, but the ratio of net capital formation
as a proportion of incomewas found to be 5.5% in the smallestfarm-
size group (0 to 2 acres)and 19.3% on the larger farms (8 acresand
above).For unirrigatedvillages,the correspondingfigureswere lower
-2.6% on the smallestfarms, and 11.2% on the larger ones.
It follows that a reduction in concentrationof land will reduce the
averagesavingsrate of the farm sector.But, again,if a compensatory
increasein total income can be securedby intensifying inputs per
unit of land soonafter land reform, the aggregatesavingscanbe pre-
vented from falling. This addsto the urgencyof introducing effective
agrarian reform (including improved technology and services)along
with land reform.
A policy implication, from the foregoing,is that the farm-sizestruc-
ture created by any land reform program should fix a minimum as
well as a maximum farm size.The minimum farm size clearlyshould

33
be determinedon the basisof the current national norm of minimum
family income. But one of the criteria for determining the minimum
income itself should be that it should at leastenablethe smallholder
to ceaseto be a "dissaver."An analogouscriterion can also be de-
rived from the known behaviorof marketedsurplus:the smallholder
should haveat leastenoughlandfor positivesales.

Tenancy Reform
The most successfulland reforms include those whereby tenants
become owners of the land they operate, as in Japan,Taiwan and
somepartsof Europe.Ownership control and incomefrom the land is
thus redistributed. However, if landlords are allowed to retain land
that might be self-operated,and tenantsbecomeowners of the land
that they operate, then the size distribution of operational holdings
maynot change.With the conversionof tenantsinto owners,security
of tenure is greaterand incomesfor the farmers are larger. This, in
turn, encouragesincreasedsavingsand, hence, on-farm investment
and higher output.
The conversionof tenantsinto owner-operatorsgenerallyleadsto
a more efficient and more equitable form of production organization
than tenancy. This is seen not only from the reforms in Japanand
Taiwan, but also from experience in parts of Africa where "cus-
tomary" tradition is convertedinto freehold. In Kenya,the provision
of security of tenure, especially in the temperate production areas,
hasincreasedon-farm investmentand helped raiseoutput.
There may be situationswhere tenancy reform aims at stabilizing
the position of tenantswith respectto rent paid, security of tenure
and labor objectives,without transferringownership rightsto them.
Here,the problem is to promote more efficient typesof tenancy,with
contracts having well-defined incentivesand deterrents.The expert
consensusis that fixed cash-rentcontractsare superior to the more
common crop-sharecontracts,since the whole income in excessof'
the fixed rentaccruesto the actual cultivator.Sharecroppers,
however,
often havea preferencefor crop sharingbecauseit provides risk in-
surance.Crop sharing can be made more efficient and equitable if if
is consideredwith cost sharing.There is growing evidencefrom the
Philippines, for example, that since the seed-fertilizer technology
began to spread, landlords and sharecroppershave spontaneously
begun trying to combine cost sharingwith crop sharingbecausethe
combinationis profitable to both.
Tenurial reforms, whether through the distribution of the land to
those working it or the provision of greatersecurity of tenure and

34
improved rental contracts, have an effect on development. Such
reforms improve income distribution by shifting income away from
the landlordsto small-scaleproducers,often thoseamong the lowest
income groups. The more secure producers tend to invest part of
their higher earningsin their holdings-thus raisingthe level of invest-
ment in agricultural production-whereas absentee landlords fre-
quently invest in off-farm activities. Finally,greatersecurity enables
tenants to benefit from appropriate technological changes,instead
of being displacedwhen landlords find it to their advantageto adopt
a different technology. The financial returns to the landlord from
using machinesand hired labor may be high, but the returns to the
economy are usually higher from labor-intensiveoperations under-
takenby smallholders.

Implementation Issues
If reforms are to generatethe benefits expected of them, several
important considerationsmust be taken into account. Firstly, since
agriculture is a private sector activity in most countries, production
and investmentdecisionsaremadebymillionsof individualsoperating
in their own interests.Very often the greaterpart of national output
comes from medium-scalefarmers.These farmers,like prudent in-
vestors,weigh the risksas they perceivethem before makingon-farm
investments-the major component of total investmentin agriculture.
Sustaineduncertainty about a government'sintentions with regard
to the distribution of land adds to the risk of investmentand can
hamper capital formation and production. In some instances,con-
tinued uncertainty hasled to disinvestmentin agriculture by owner-
operatorsand a flight of capital from the country. It follows that the
more specific the plans and the more clearly defined the policies
regarding land reform, the less likely the accelerationof disinvest-
ment by landownersand, so,the lower the "cost" of the reform.

LogisticalSupport
Secondly,the introduction of a major land reform programusually
disrupts the systemof logistical support from the commercialsector
to the farmers. In most countries in the world, there is a well-estab-
lished link between commercial bankersand suppliersin the private
sector and the larger agricultural producers.This linkage is basedon
mutual interests and, often, on long-standingbusinessassociation.
The redistribution of land frequently leadsto a breakdown of this
system.Often, there is a long interval before the public sector can

35
undertake the role previously filled by the private sector, or before
the privatesectoradjuststo the new situation.Without an appropriate
organizationfor the provision of inputs, productivity will decline and
output will fall. Thus,the reduction of the costsof a land reform pro-
gram-in terms of production forgone-depends on the rapid reor-
ganizationof the input supplysystem.

Natureof Organizations
Thirdly, the natureof the organizationsproviding for both the sup-
ply of necessaryinputs and the marketingof production surplusesis
crucial in a post-reform period. There are many different forms of
organization: cooperatives,agricultural development banks,special
credit institutions, marketing authorities,and the like. Whatever the
organizationsthat prevail, it is essentialthat they be designedspe-
cifically to assistthe beneficiariesof reform. In many instances,the
institutions that have provided servicesin a post-reform period have
continued with a bias in favor of larger-sizeoperations. Part of the
reason is that these institutions have not been able to adapt their
methodsof operationto the needsof largenumbersof small farmers.
Unlessthis is done, the beneficiariesof the reform may not be in a
position to increasetheir output. Indeed,the appropriate organiza-
tion of supplies and the evolution of a low-cost delivery systemto
reachsmall-scaleproducersis a sine qua non for a sustainedincrease
in productivity.

Adaptation
Fourthly, under certain conditions land reform programs might
need adaptation if they are to fulfill the objectivesof development.
When land is fully utilized and yields are alreadyhigh, the impact of
redistribution of land on productivity and employment may be in
question. In this context, it is important to determine the reasonsfor
high yields. In much of agriculture, mostof the inputs are "divisible,"
thus reducing the importance of scale of operations as a factor in
raising productivity. In somesituations, high yields and efficient op-
erations may be directly associatedwith a systemorganizedto func-
tion on a large scale (as in certain types of sugar plantations).The
breaking up of such holdings may well reduceyields and lower out-
put. A more realistic approach to obtaining widespread benefits
would be to leavesuch operationsintact and redistributethe profits
from the enterprise.Thiscan be done through taxation, by raisingthe
wagesof the workers, or-as in Peru-converting the operation into

36
a worker-owned corporation and distributing dividends, out of
profits, to the participatingstockholders.

Structural
Change
Finally,land reform leadsto structural changeswithin the agricul-
tural sector. The post-reform structure will depend on the ideology
of the government. In some instances,the number of small-scale
owner operationswill increase;in others, producer cooperativesor
communesor large-scalestate farms will emerge. The pattern that
evolves may also be tailored to fit the economic environment: the
organizationmight be basedon a systemwhich canusesurpluslabor
for direct capital formation; other organizations(suchas large-scale
state farms) might be intended to save labor. Experiencehas indi-
cated,however,that:
1. Government reorganizationcan generateenthusiasmand pro-
vide opportunitiesfor mobilizingworkers, but raisingoutput depends
on more than land and labor. There must be an appropriate supply
of other inputs.
2. No matter what the structure,an appropriate systemof manage-
ment is necessarywhich enablesthe managersof land to makedeci-
sionsin a timely fashion-a most important condition in agriculture
and one that is dependenton weather.This is a condition, however,
that is often unfulfilled in rigidly controlled societies.
3. There must be an adequatesystemof incentivesand rewardsif
productivity in agriculture is to be increased.Thisappliesboth to the
agriculturalsector as a whole and to the units in which beneficiaries
of reforms are organized. Many communes,producer cooperatives
and other units of production haveflounderedin developingasystem
that reflects both equity and incentives.The creation of adequatein-
centives is particularly important in a situation where labor is the
major input.
Land reform, although equity oriented, can be consistentwith all
the goals of economic development: raisingproductivity, increasing
employmentand providing wider equity. In the long run, land reform
need not leadto a reduction in marketedoutput or savings.Tenancy
reformscanredistribute incomesand, by providing securityof tenure,
can encourage increasedon-farm investment. However, sustained
increasesin output dependon complementaryinvestmentsand poli-
cies.The most important of theseconcernthe organizationand pro-
vision of an adequatesupply of inputs for the beneficiariesand the
creation of incentivesto use theseinputs to raiseproduction.

37
 Chapter 3: THE WORLD BANK AND
LAND REFORM
ChangingConcerns
The position of the World Bank in regard to land reform has
changed over the past decade, reflecting a reconsiderationof the
objectives of development and the most appropriate strategiesfor
attaining thoseobjectives.Theobjectivesarenow generallyaccepted
to be increasedproductivity and employment,and socialjustice. Land
reform canbe corsistentwith theseobjectivesand, in somesituations,
maywell be a necessarycondition for their realization.
In the early yearsof the Bank'soperations,the focus was on pro-
viding adequateinfrastructurefor increasingagriculturalproduction.
In the early 1960s,the approach to agricultural development was
widened to include the provision of rural credit and on-farm inputs.
Problemsof tenurewere seento havean indirect bearingon produc-
tion, mainly becausethey influenced on-farm investment decisions
and determined the efficiency of resourceuse, especially irrigation
water. By the end of the 1960s,however, concernwas growing about
distribution of income in the rural areasand the relationshipbetween
land distribution and income distribution. This was reflected in the
Agriculture SectorWorking Paperof June1972,which recognizeda
relationship between land distribution and equity. The paper stated:
"In developingcountries, landrepresents a muchhigherproportionof
totalwealththanin developedcountries, and inegalitarian
patternsof
landownership area majorsourceof incomeinequality.Furthermore,
the ownersof landusuallypossess politicalandeconomicpowerwhich
can be exercisedin waysthat harmthe interestsof the bulk of the
ruralpeople."
The paperwent on to affirm that:
"It is clearthatagriculturaldevelopment cannotdo all it mightto im-
proverurallife if the distributionof landownershipis highlyskewed."
This concern has been reflected both in the technical assistance
offered to governments (especiallyin sector survey and economic
reports) and in the types and componentsof projects in the lending
program.

TechnicalAssistance
The Bank hasbeen concernedwith problems associatedwith land
distribution and land reform since the beginning of its operations.
One of the first major economic surveys undertaken was that of
Colombia in 1955.The missionidentified the patternsof landuseand

38
land distribution by sizeof holding to be major obstaclesto accelerat-
ing agriculturaldevelopment.Largestretchesof fertile landwere held
by large-scaleproducersfor livestockraising,while intensiveagricul-
ture was practiced by "minifundios" on land that was lesssuited for
crop production. The mission recommendedto the governmentthat
it introduce a graduatedland tax as a meansof intensifying land use.
A subsequentagriculture sector mission in 1956 confirmed that the
systemsof land tenure and land usewere barriersto increasingout-
put. This mission recommendedthat the government adopt a pre-
sumptive income tax to encouragethe more productive useof land.
The two missionsto Colombia were concerned with increasing
productivity and intensifying land use. The missionswere not con-
cerned with the redistribution of land as a means of encouraging
greater equity, nor did they consider redistribution as a meansof
intensifyingproduction. Rather,they took the view that the distribu-
tion of land was a matter of national policy and internal politics, and
that the Bank-as an external lending agency-should adhere to the
existingpolicy and not advocatea rapid redistribution of land. It did,
however, recommenda vigorous policy of settlementon reclaimed
and clearedland.
Since that time, missionsand sector surveyshavebeen conducted
in almost all the countries servedby the Bank.Many of these have
pointed to patternsof land control and insecurityof tenure as obsta-
cles to raisingagricultural productivity.More recently,there hasbeen
a growing emphasison the problems of distribution of land and the
rightsto land as factors that influence equity aswell as productivity.
Thus, missionsto Ethiopiaand Morocco havedrawn attention to the
relationship between the land tenure situation and the distribution
of benefits from growth. In Morocco, the mission emphasizedthe
possibility of redistributing landas a meansof increasingboth output
and equity. In Ethiopia,the problem was seenas one of unevenland
distribution and insecurity of tenure; securityof tenure was consid-
=ered to be especially significant in the light of the distribution of
potential gainsfrom new technologybeing introduced into the coun-
try. Landlordswere finding it increasinglyprofitable to displacetheir
tenantsas machinetechnology provided higher returns.
Despitethis trend, many reportsdo not give appropriate emphasis
to issuesrelated to land reform and development.The Bank needs
to be better informed about conditions governing rights to land and
related institutions in member countries. More needsto be known
about the distribution of land,conditions governingtenancy,and the
policies and programsinstituted to influencethe distribution of land
and rural incomes.it is only through a thorough analysisof conditions

39
within member countriesthat the Bankwill be in a position to discuss
policy options with member governments.At present,many reports
still do not addressthese problems; however, new guidelines are
being developedwhich can form a basisfor discussingthe issuesin
a systematicway in sector and economic reports.

Lending Operations
The Bank'slending for agriculturaldevelopmenthasincreasedvery
rapidly in recentyears.Loansand credits havebeen madeto countries
with widely differing social and political structures.These have in-
cluded socialistcountries,such as Yugoslaviaand Tanzania,as well
as countries that follow capitalism,such as Argentina and Thailand.
Loansand credits have been made for agriculture operating under
different forms of tenure-for kombinatsin Yugoslavia,kibbutzesin
Israel, individual holdings in India, cooperativeproduction units in
Tunisia and group farmersin Kenya.Fundshavealso been provided
for large-scalelivestockproducers,large-scaleplantationsand small-
scaleproducers;thesehavebenefited absenteelandlords,large land-
owners, small landowners,tenants and farm workers. On the other
hand, the Bank has not been totally indifferent to structural and in-
come distribution aspects,and the recordshowsan increasingaware-
nessof the implicationsreflected in more frequent useof measuresto
improve them.
Nevertheless,few projects havesupported land reform as such. In
general, external financing, whether multilateral or bilateral, has
played a minor role in the financing of land reform programs.One
reason is that the processof reform in itself may only require rela-
tively small outlaysof public funds,asexpendituresfora redistributive
reform depend mostlyon the levelsand forms of compensationthat
are set for the former landowners.Public discussionof land reform
financing is generally dominated by this issue.When land is confis-
cated as part of a revolutionary process-as it was in Mexico and
Bolivia-clearly little, if any, public expenditure is involved. The
compensationissuetendsto be more important in such countries as
Colombia and Venezuelawhere land is purchased.Evenso,the actual
amountsinvolved are not substantial,especiallywhere, as is usually
the case,paymentis mostly in bonds. It is estimatedthat, in the Latin
American countries which followed nonconfiscatoryreforms, only
some9% to 15% of total reform-relatedcashbudgetswent for land-
owner compensation-though in other casesthe figure could be
muchhigher.
Compensationpaid for land is a "transfer payment" from the pub-

40
lic sector to the landholding groups.Without doubt, compensation
can haveserious implications for income distribution, consumption
and investment-but it does not of itself create any new productive
capabilitiesin the country. Partlybecauseof this, internationallending
institutions have refrained from using their resourcesfor financing
land purchases.It hasbeen suggestedthat the internationalagencies
might guaranteebonds issuedto compensatelandlords.If financing
were to be through international maintenance-of-valueguarantees
of bonds and for compensation,this would have the paradoxical
effect of giving land bonds greaterstability than that enjoyed by the
currenciesof issuingcountries.
The Bankhasprovided generalsupportfor at leastone far-reaching
land reform program.This was in Tunisia where the Bankprovided
a loan of $18 million intended to back a major agrarianreform relat-
ing to former French-ownedestates,which occupied the most fertile
land in that country. The nationalizedland was to be converted into
"units of production" which were to be farmed on a cooperative
basis;eachunit of production was to be self-financingand, inter alia,
was to pay a guaranteedminimum cashwage to the workers out of
the farm profits. However,the scarcityof trained manpowerand the
rapid paceadoptedin establishingnew cooperativesmade it difficult
for the production units to start on a sound basisand generatea large
enoughcashflow to meet their objectives.In addition, the systemhad
built-in disincentivesbecausewageswere not paid accordingto work.
The Bank successfullypressedfor substantial improvementsin the
conception,designand implementationof the agrarianreform. It was
unable, however, to influence the major political decision either to
take all the land in Tunisia under state managementor to put it all
under the control of cooperatives.The extensionof reform strained
the limited administrative capacity, and the reform program col-
lapsed.Smallholdersopted for private farming and were supported
by landownerswho resistedthe takeover of their lands. The Bank
subsequentlycanceledhalf of the loan.
The problems encountered in financing the Tunisian program
underscore some of the difficulties in lending for reform-related
projects. The financial viability of these projects dependsto a great
extent on the managerialcapacityof the beneficiariesof the reform
and the development of an efficient service systemfor them. Very
often the managerialcapacityof the beneficiariesmaybe untried; the
agenciescreatedto deliver the inputs are usuallynew, have limited
technical capacityand are of questionablefinancialviability. Further-
more, these institutionsoften provide inputs that were formerly pro-
vided by the private sector, and the whole delivery systemchanges

41
 from one basedon the profit motive to one basedin the first instance
on social consideration.This directly affects their financial viability,
especiallyin that cashflows generatedby reform projects tend to be
lessimmediatethan in other projects,and many investmentsin social
overhead are not self-liquidating in the short run.
Another Bankproject provided direct financial assistanceto facili-
tate the implementation of land reform as part of the Lilongwedevel-
opment schemein Malawi. It was recognizedduring the preparation
of the Lilongwe project that there was an opportunity to changethe
existing land tenure pattern of customaryright of usufruct.The need
for changeto a more secureand lastingtenure systemwas evident
asalmost all uncultivatedlandhad been takenup; individual holdings
were of the order of about five acresper family, and fragmentation
of holdingshad occurredon a substantialscale.Fiveacreswasdeemed
to be the minimum holding size capable of providing a family with
subsistenceat presentlevelsof technology.
As a consequence,the Malawi Governmentintroduced three Acts
of Parliamentwhich provided for the allocation, consolidation and
registrationof holdings,and the issuanceof either family or individual
freehold titles. These Acts also provided for the regulation of the
subsequentsale,mortgageor transfer of registeredland through the
establishmentof LandBoards.To date, some200,000acreshavebeen
allocatedand titles issuedon 60,000acres.IDA creditsare beingused
for the land survey (both topographicaland cadastral),the provision
of allocation and registrationstaff,vehicles,equipment, and the con-
struction of housing and land registry.The amount involved will be
approximately US$1 million by the end of the second phase.The
Lilongwe project indicates that Bank assistancecan play a role in
assistinggovernmentsin the "mechanics" of land reform and in the
draftingof legislation.
A number of other projects have been financed by the Bank in-
volving somechangein distribution of landor in tenurial rightswithin
the areaencompassedby the project. Theseinclude projectsfor land
settlement,outgrower schemes,irrigation, and rural credit.

LandSettlement
The Bank hasfinanced a number of settlement projects in which
infrastructure was made available together with other servicesfor
families settled in the project area.Table 3 gives information on ten
projects located in Brazil, Colombia, Ethiopia, Kenya, Malawi and
Malaysia.Sevenof the projects were establishedon public land and
so did not involve any change in the size distribution of existing

42
 Table 3
Costs of Selected Settlement Projects Assisted by the World Bank
Estimated
Total Bank or IDA finance Number of project Average
project families(') costs per farm Settlement
costs Amnunt Lnanor to be family(') size on
Country Project (US$ millions) (US$ millions) credit Date settled (US$) (hectares)
Brazil Alto Turi Land
SettlementProject 12.6 6.7 loan 1972 5,200 2,423(3) 40.0 Publicland
Colombia AtlanticoNo. 3
Irrigation 15.7 9.0 loan 1967 2,500 6,280(4) 4.0(5) INCORAland
SecondAtlantico (involved
Development 9.7 5.0 loan 1972 1,800 5,389 11.0 appropriation
CaquetaLand land)
Colonization 21.6 8.1 loan 1971 6,300(6t 3,429 n.a.(7 l
Ethiopia WolamoAgricultural
Project 2.3(8) 3.5 credit 1969 1,050 2,214 6.0 Publicland
t Kenya LandSettlement European-
and Development 6.9 3.9 credit 1969 5,200 1,327 14.3 ownedland
Malawi Karonga
RuralDevelopment 7.8 6.6 credit 1972 2,830 2,756 6.0 Publicland
Malaysia JengkaTriangle 29.1 14.0 loan 1968 2,770 10,505 4.8 Publicland
Second
JengkaTriangle 41.0 13.0 loan 1970 3,000 13,667 4.3 Publicland
Third
JengkaTriangle 43.3 25.0 loan 1973 4,000 10,825 4.5 Publicland
Source: World Bank and IDA appraisal reports.
(l) Except for Kenya, figures represent goals rather than actual state of settlement.
(,) Project costs, as estimated in the appraisal reports, do not necessarily reflect total economic costsof settlement.
(3) The costtothe government is$1,7nnperfamily settled.This excludes expenditureson health, education, research and related studies. Thesecost expenditures are being reviewed and are expected
to be Considerablyhigher than originally expected.
)') The cost per small farmer settled is estimated to be $17,000, whereas the cost per middle-size farmer remaining in the project area is $100,000.
(5) The original goal was to settle 2,500 landless peasants and develop 9,900 hectares. The project is behind schedule.
(6) Includes 2,800 now settlers and 3,500 partially established settlers.
(a) Although 2,800 new settler families are scheduled to be settled on some 280,000 hectares, no data on the farm size of 3,500 partially established settlers are given.
(8) Excludes $2.73 million used for agricultural development on the highlands.
holdings.Thus,settlerswere allocated holdingsof from three or four
hectaresin Malaysiato 40 hectaresin Brazil.Eachholdingwasdeemed
adequatetcprovide a livelihood and full employmentfor the settler
and his family.
There are severelimitations on settlementas a meansof reaching
large numbersof landlesspeople or relieving pressureson the land.
Although the costsper family in a settlementproject can be mislead-
ing, the data in Table 3 indicatethe limitationson settlementprojects
-as presentlyconceived.Theten projects were intended to settle no
more than 35,000families; the total cost was expected to be $190
million, the Bank'scontributions being almost half that amount.The
capital requirementof more than $5,000per family limits the pros-
pects of the approach.Clearly,the whole approachto capital-inten-
sive settlementrequiresreexaminationconsideringthe magnitudeof
the problem outlined in Annex 1 of this paper.

OutgrowerSchemes
The problems of distributing the gains from plantation develop-
ment were mentioned earlier. It was suggestedthat the benefits be
distributed through the raisingof wagesand the paymentof dividends
to the workers. In this area, the Bank has made a substantialcon-
tribution toward a novel form of tenure through the developmentof
"outgrower" schemes.Theseschemesinvolve the production of tree
crops on smallholdings rather than on large-scaleplantations. The
smallholdings are establishedaround the nucleus of either a pro-
cessingplant or a plantation.Thecentralunitprovides technicalassist-
ance, inputs and marketingservicesfor the outgrowerswho, in turn,
sell their productsthrough the centralorganization.
TheBankhasparticipatedin ninesuchprojectscosting$125million,
of which the Bank has contributed $68 million and affecting some
120,000families.Thesehaveincluded teaprojects in Indonesia,Kenya,
Mauritius and Uganda, rubber in Indonesiaand Malaysia,cocoa in-
the Ivory Coast,and oil palm in Nigeria.The averageholding in each
project hasrangedfrom 10 hectaresin Senegalto one acre in Kenya.
In the main, the size of holdings for outgrowers is small, although-
large enough, under labor-intensive cropping systems,to employ a
family and produce enough of a high unit value commodity to yield
an income well in excessof that earnedby producersof staple com-
modities who have holdings of a similar size. While this systemhas
madea valuablecontribution toward establishingviablesmallholders,
it is only effective when there is a commodity that can be handled
througha centralprocessingsystem.

44
Irrigation
The Bank has invested about $1,450 million in irrigation, flood
control and drainage projects. While these projects covered many
facetsof water storageand distribution, most were intended to im-
prove the use of water and-bring more land under intensivecultiva-
tion. To this end, the Bank hasworked with various governmentsin
determiningthe mostappropriatesize of holding for the beneficiaries
of eachproject. For example,11 projects costing$342million (incor-
porating a Bankinvestmentof $190 million) are expectedto improve
810,000hectaresand benefit more than 500,000families.The average
size of holdings in the irrigated areasrangesfrom 10 hectaresin Iraq
to one hectarein Korea, Pakistanand Sri Lanka,or an averageof 1.6
hectaresper family over alI the projects.
In many instances,irrigation projects are subjectto special regula-
tions or laws regarding the size of holding that can be held by the
beneficiary.Thus, in Mexico the Bank-supportedprojects have con-
formed to the law which limits the size of irrigated holdings to a
maximum of 10 hectares.Elsewhere,problems have arisen because
there is no legal provision regardingsize of holding or becausethe
law has been ignored. In some instances,the Bankhas insistedon
special legislation giving tenantssecurityof tenure. But, in practice,
this hasbeendifficult to enforce.

RuralCredit
While in itself farm credit is an important instrumentfor reaching
groups of a particular size in agriculture, accesscan be restrictedby
tenurial arrangementsif lending criteria specify that registeredland
titles be usedascollateral for borrowing. Bank-assisted projects have
provided more than $1,000 million for rural credit. Most of these
resourceshaveaided largercommercialproducers,although in recent
yearsthere hasbeen a pronouncedtrend toward lending for smaller
producers.By the end of 1973,an estimated$250 million had been
allocatedfor small farmers.
In some instances,the Bankhasmade loans on the condition that
the recipient government takes steps to ensure that the intended
beneficiariesdo indeed gain from the investment.However, in sev-
eral instances,the governmentsconcernedhave not fulfilled obliga-
tions regardingthe provision of securityfor tenantsor the allocation
of land to low-income groups. In other instances,governmentshave
failed to implementconditions provided for by existinglegislationon
rights to land; or they have failed to introduce legislation which
would havemet the conditions specified in the loans.Thishighlights

45
one of the major dilemmas confronting an international lending
agencyconcerned with promotion of land reform as an instrument
of economicdevelopment.That is, to what extent canthe Bankinflu-
ence the courseof eventsregardingdistribution of land, and income
from the land, in the sovereignstatesthat are membersof the Bank?

Major Policy Options

The Bankhasto recognizethat its leverageis limited as it seeksto
redefineits positionwith regardto land reform.UsingBankfinanceto
gain a developmental impact through land reform involves highly
complex issuesat the project level, while the potential for usingthe
Bank'sinfluence to pressor even force the issueof structural reform
on member countries is severelycircumscribed.Suchpolitical deci-
sionsare not amenableto ready negotiationwith governmentsin the
sameway as are other institutional questions-such as, for instance,
the settingof public utility rates.
The Bankwould seemto be left with only two options. Firstly,in
countriesthat are interestedin pursuingland reform the Bankcangive
support in the form of technical assistanceand finance for reform-
related projects. It should give overt priority in lending to those
countries and projects which meet land reform criteria. Secondly,in
countries where governmentsare not interested in land reform the
Bankshould:
(1) studythe situation in all cases;
(2) call the attention of the governmentsto the problemsassociated
with the existing tenure system,and enter into a dialogue on the
subject;
(3) support land reform proposalswhen they are made officially;
and
(4) not lend for projects if tenurial arrangementsare so bad that
they frustrate the achievementof the Bank'sobjectives.
These options are reflected in the policy guidelines provided in
this paper.

46
ANNEXES
I I
I
 Annex1

THE CONTEXT OF
LAND REFORM
Ratios of Population to Land
The total land area of the globe is about 13,393million hectares,
madeup of 1,456 million hectaresof cropland,defined as arableland
and land under permanent crops (10.8%); 2,987 million hectares
under permanent pasturage (22.8%); and 4,041 million hectares
under other uses(36.4%).Of the arable land, approximately32% is
in Asia; 19% in North and Central America; 16% in the USSR;15%
in Africa; 10% in Europe;6% in SouthAmerica; and 3% in Oceania.
The world's population was estimated at approximately 3,617
million in the early 1970s.This representsan averageof 3.7 hectares
of land, or closeto 0.40 hectareof cropland, per person.Theworld's
agriculturalpopulation-defined as populationdependingon agricul-
ture for its livelihood-is estimatedat 1,851 million, or 51% of the
total population. On the basisof these global figures, there is an
averageof 0.78hectareof cropland per person in agriculture.
The relationship between population and land in all major regions
and for 52 selectedcountries is shown in Annex Tables1:1 and 1:2,
respectively.Among other things, the tables show that:
1. More than 70% of all rural people live in Asia,which hasapprox-
imately 32% of the world's cropland. The ratio of cropland to agri-
cultural population is the lowest in Asia among all the major regions,
averaging0.35 hectare per person.Together,the People'sRepublic
Table 1:1
Regional Distribution of Land, Cropland, Agricultural
Populationand Area per Personin Agriculture
Cropland Ruralpopulation Agricultural
population as Croplandarea
Landarea Distri- Distri- percentage of per rural
(million (million bution bution total popu- person
Region hectares) hectares (%) (millions) (%) lation (hectares)

Europe 493 145 10.0 89 4.8 17 1.63

USSR 2,240 232 15.9 77 4.2 32 3.01
Northand
CentralAmerica 2,242 271 18.6 54 2.9 17 5.02
SouthAmerica 1,783 84 5.8 74 4.0 39 1.14
Asia 2,753 463 31.8 1,314 71.0 64 0.35
Africa 3,031 214 14.7 239 12.9 67 0.90
Oceania 851 47 3.2 4 0.2 4 11.75
Total 13,393 1,456 100.0 1,851 100.0 51 0.78

Source:FAO.ProductionYearbook
1972.

49
 Annex 1
Table 1:2
Cropland in Relation to Population, by Country
Hectaresof cropland
Total Agricultural per personof:
Cropland population population Total Agricultural
Country (000hectares) (000) (000) population population
Africa
Angola 900 5,501 3,568 0.16 0.25
Ghana 2,835 8,832 4,840 0.29 0.59
Ivory Coast 8,859 4,916 3,986 1.80 2.22
Nigeria 21,795 76,795 45,423 0.32 0.48
Rwanda 704 3,609 3,277 0.20 0.21
Uganda 4,888 8,549 7,342 0.57 0.67
Zaire 7,200 17,493 13,701 0.41 0.53
Asia
Bangladesh 9,500 71,000 60,000 0.13 0.16
Burma 18,941 27,584 17,570 0.69 1.08
China,People'sRepublicof 110,300 850,406 568,921 0.13 0.19
China,Republicof 867 14,520 6,171 0.06 0.14
India 164,610 550,376 372,605 0.30 0.44
Indonesia 18,000 119,913 83,230 0.15 0.22
Japan 5,510 103,540 21,329 0.05 0.26
Korea,Democratic Republicof 1,894 13,674 7,275 0.14 0.26
Korea,Republicof 2,311 32,422 17,300 0.07 0.13
Malaysia 3,524 10,931 6,176 0.32 0.57
Nepal 2,090 11,040 10,112 0.19 0.21
Pakistan 24,000 60,000 35,000 0.40 0.69
Philippines 8,977 38,493 26,752 0.23 0.34
Thailand 11,415 35,814 27,398 0.32 0.42
Viet-Nam,Democratic
Republicof 2,018 20,757 16,108 0.10 0.13
Viet-Nam,Republicof 2,918 18,332 13,620 0.16 0.21
Europe
Denmark 2,678 4,921 595 0.54 4.50
GermanDemocratic Republic 4,806 17,257 2,133 0.28 2.25
Germany,FederalRepublicof 8,075 61,682 3,514 0.13 2.30
Hungary 5,594 10,310 2,484 0.54 2.25
Italy 14,930 53,667 9,735 0.28 1.53
Poland 15,326 32,805 9,940 0.47 1.54
Portugal 4,370 9,630 3,523 0.45 1.24
Romania 10,512 20,253 10,503 0.52 1.00
Spain 20,601 33,290 11,222 0.62 1.84
Sweden 3,053 8,046 754 0.38 4.05
United Kingdom 7,261 55,711 1,540 0.13 4.71
USSR 232,809 242,768 77,322 0.96 3.01
Yugoslavia 8,205 20,527 9,651 0.40 0.85
Latin America
Argentina 26,028 24,353 3,704 1.07 7.03
Bolivia 3,091 4,931 2,873 0.63 1.08
Brazil 29,760 93,565 40,869 0.32 0.73
Chile 4,632 9,780 2,484 0.47 1.86
Colombia 5,258 21,117 9,541 0.25 0.55
Cuba 3,585 8,407 2,755 0.43 1.30
Guatemala 1,498 5,180 3,246 0.29 0.46
Haiti 370 4,867 3,754 0.08 0.10
Mexico 23,817 50,670 23,617 0.47 1.01
Peru 2,843 13,586 6,189 0.21 0.46
PuertoRico 236 2,784 387 0.09 0.61
Uruguay 1,947 2,886 482 0.67 4.04
Venezuela 5,214 10,997 2,887 0.47 1.81
North America
Canada 43,404 21,406 1,712 2.03 25.4
UnitedStates 176,440 205,395 8,216 0.86 21.5
Oceania
Australia 44,610 12,552 1,049 3.55 42.53
Source:Dovring,Folke.landReform:
EndsandMeans.
A Background
Studypreparedfor the WorldBank.
 Annex1
of China and India havean agricultural population of close to 1,000
million, while Indonesia,Bangladeshand Pakistanhavea further 178
million. Of the Asian countries, in terms of hectares per person,
Burma hasthe most favorable ratio of cropland to rural population
(1.08),followed by Pakistan(0.69),Malaysia(0.57)and India (0.44),
comparedwith Indonesia(0.22),the People'sRepublicof China(0.19)
and Bangladesh(0.16).The leastfavorable ratio is in the Republicof
Korea and the Democratic Republicof Viet-Nam (eachwith an esti-
mated 0.13). It is notable that the Republic of China (Taiwan)and
Japanhave ratiosof 0.14 and 0.26arable hectaresper person in agri-
culture. Japanis the only developedcountry with such a low ratio-
well below the 1.63of Europeand 5.02of North and CentralAmerica.
2. SouthAmericaaccountsfor 4% of the world's agriculturalpopu-
lation and 5.8% of the world's cropland. Although 13% of the land
area of the world is in South America, almost half of that area is in
forests and woodlands, 20% is in pasturelandand only 5% or 6%
is in cropland. However,as only 39% of the population is in agricul-
ture, there is an averageof 1.14 hectaresof arable land per rural
person.Argentinaand Uruguay have high ratios of agricultural land
to rural population, the most favorablein the developingworld (7.03
and 4.04, respectively).Venezuela,Chile, Bolivia, Mexico and Cuba
have ratios of more than 1 hectareper person in agriculture; Brazil,
Colombia, Peru and the crowded Central American republics have
ratios of lessthan 1 hectareper rural person.Haiti with 0.10 hectare
per person in agricultureappearsto have the most unfavorableratio
in the world.
3. Africa has13% of the world's rural population and closeto 15%
of the world's cropland, with an averageof 0.90hectareof cropland
per person in agriculture; 67% of the population dependson agricul-
ture, a higher proportion than in anyother region.The mostfavorable
ratio in tropical Africa appearsto be in the Ivory Coast,with 2.22
hectaresper person in agriculture.Uganda,Ghana,Nigeriaand Zaire
have between 0.50 hectareand 0.70 hectare per person-in agricul-
ture. Rwanda,with 0.21 hectareper person in agriculture, is one of
the few countries in tropical Africa where the pressureon land re-
sourcesis greaterthan the averagein Asia.
This brief summaryindicatesthe wide rangeof population densi-
ties in rural areasin different regionsand countriesof the developing
world. The data show that, by and large, countrieswith a high pro-
portion of population in agriculture have less favorable ratios of
population to land. They are also among the poorest countries.Fur-
ther, they are the countries in which population is increasingrapidly
and where it is particularly difficult to raiseagricultural output.

51
 Annex 1

PopulationandProduction
The population in the rural areasof developing countries,while
declining relative to total population, is increasingin absolute num-
bers. Despite rapid migration out of agriculture, and despite the
explosivegrowth of population in certainareas,the rate of growth of
the rural population has increasedin all regions of the world other
than Africa. Table 1:3 showsthe trends in rates of growth between
1950-60and 1960-70,with overall growth rates rising from 1.9% to
2.1%, and the largest regional rate of increasebeing the one from
1.8% to 2.1% in EastAsia (where population density is alreadygreat
in rural areas).
Table 1:3
Rural Population Growth, by Region
Annual percentage rate
1950-60 1960-70

Latin America 1.4 1.5

EastAsia 1.8 2.1
MiddleEast 1.8 1.8
Africa 2.4 2.2
Totalall regions 1.9 2.1

Source: Davis, Kingsley. WorldUrbanization,1960-70. Vol. 1,1969.

The larger number of people hasadded to the pressureof popula-

tion on the land. Historically,this pressurehasbeen relievedthrough
the expansionof acreagealong a frontier of cultivation. Indeed,it was
the expansion of the frontier in the new lands of North America,
Argentina, SouthAfrica and Australiathat helped relieve population
pressuresin the first period of generalizedpopulation growth in the
late eighteenthcentury. In theseareas,population growth was accel-
erated by an influx of migrants to rates comparable to those found
today in many of the poorer countries. However, since the frontier
is fast disappearingin mostof the poorer countries,so arethe oppor-
tunities for low-cost expansion of acreage under cultivation. The
changingsituation is difficult to document at an aggregatelevel, but
Table 1:4 gives some perspectiveson trends in the expansion of
cropped areasand production.
The rate of expansionin acreagefell, in the aggregate,in the 1950s
and the 1960s.The only exceptionis LatinAmericawhere the acreage
under cultivation grew from a rate of 1.8% to 2.5% per year. In all
other areas,the expansionof acreageslowed down, halving in the

52
 Annex 1
Table 1:4
Cropped Area and Production Trends, by Region
Averageannualgrowthrate
1953-55to 1962-63 1961-63
to 1969-71
Production Area Production Area

LatinAmerica 3.1 1.8 2.9 2.5

EastAsia 2.5 1.9 2.8 1.1
MiddleEast 3.8 2.2 2.7 1.1
Africa 3.0 1.7 2.6 1.2
All regions 2.8 1.9 2.8 1.4

Source:FAO.Reportanthe
1960Wo1id
CensusofAgricalture.
Rome:1971.

Middle Eastfrom 2.2% per year to 1.1%. When the rates of popula-
tion growth are compared with rates of increasein acreageunder
cultivation, it appearsthat the rural population increasedat about
the same rate as the cropped area during the 1950s,but increased
more than one-and-a-halftimes as fast as the cropped area during
the 1960s.
As shown in Table 1:4, production increasedat the sameratedur-
ing the 1950sas during the 1960s.A rate of increasein output con-
sistentwith an increasein rural population indicatesa decline in the
rate of growth of output and incomes from 0.9% per year in the
1950sto 0.7% per year in the 1960s.At the sametime, asaverageper
capita income was increasingat a declining rate, yields per acre rose
very moderately-in this instance,an increaseof around 0.4% a year
in the 1950sand 1960s.
The increasein population and slow expansionof the area under
cultivation have caused a deterioration in man-land ratios. This
deterioration, arising from constraintson the low-cost expansionof
acreageunder cultivation, makesit increasinglydifficult to accelerate
growth rates of output and income in agriculture. This is because
raisingyields requiresa higher level of technologyand management
as comparedto increasingoutput or expandingacreageunder culti-
vation. It is only in recent years that a concerted effort has been
made to develop technologiesto raiseyields of staple crops grown
in the developing areas.Hitherto, these efforts have been confined
to a handful of crops, and the successesattained havebeen limited
to a relatively small areaof the developingworld. In somefortunate
countries,such as Nigeria, someland resourcesare still availablefor
future developmentthrough an expansionof acreageunder cultiva-
tion. But many other countries have little or no unused land, so the

53
 Annex 1
situation is correspondinglyworse. The emphasisin the latter coun-
tries will have to be placed more and more on raising yields per
hectare.
The increasingpressureof population on the land highlights the
issueof absorptivecapacityin agriculture.Most developingcountries
haveconsiderableopportunities for increasingemploymentand pro-
duction in this sector. This applies to the more densely populated
regionsas well as to others.Table 1:5 showsthe startling differences
in input of agricultural labor and output per hectarein developing
countries of Asia on the one hand, and in Japanon the other. Japan

Table 1:5
Agricultural Labor Force and Production in Selected Asian Countries, 1970
Netagri-
Agri- cultural
cultural produc-
workers tion Output
per Indices per Indices per Indices
100 Japan hectare Japan worker Japan
Country hectares = 100 (US$) = 100 (US$) = 100

Burma 48 25 71 9 148 37
India 92 48 115 15 150 38
Indonesia 224 117 283 37 126 32
KhmerRepublic 75 39 146 19 194 49
Korea,Republicof 261 136 440 58 169 43
Laos 153 80 119 16 75 19
Malaysia 74 39 366 48 492 124
Nepal 229 119 220 29 96 24
Pakistan 101 53 218 29 215 54
Philippines 113 59 178 23 158 40
Sri Lanka 107 56 286 38 266 67
Thailand 119 62 179 23 150 38
Viet-Nam,Republicof 242 126 241 32 100 25
Japan 192 100 762 100 397 100

Sources:Column1: International
LabourOffice.LabourForce
Projections.Pt. I-V. Geneva:1971.
Columns3 and 5: FAO.TheStateof FoodandAgricuture,.
p. 99. Rome:1972.

is a country of small holdings and has approximately two workers

per hectarewith an averageoutput of $397 per worker and $762 per
hectare.Severalother countries havea higher ratio of workers to the
land than Japan,while one country, Malaysia,hasa higher output per
worker in agriculture than Japan.However, the point to be empha-
sized is that if the level of labor intensityof two workers per hectare
prevailing in Japancould be attained in countries such as Pakistan

54
 Annex1
and India, the agriculturalsectorin thesetwo countriescould absorb
all the labor force expectedby 1985. Thiskind of labor intensity is not
likely to be reached,however, becauseof the small size of the irri-
gated areas in Pakistanand India and other constraintsrelated to
technology, resourcebase,land tenure and capital formation.
It is reasonablyclear that whatever is done will only partially satisfy
the ever-risingdemandfor work and income in the manydeveloping
countriesthat arefacedwith the generalproblemsof high population
growth, low incomesand increasingunemployment.With very few
exceptions,the poverty and unemploymentproblemsof the develop-
ing countriesare unlikely to haveany long-term solutionsthat would
not include a reduction in population growth, urban aswell as rural.
Nonetheless,evenif effective birth control could be introduced over-
night, specialand possiblyextraordinarymeasureswould haveto be
taken to satisfy the expandingdemand for work and income from
today'schildren. Suchmeasuresinclude thoserelated to land reform.
Distributionof Land
The ratio of population to land tells us nothing about the distribu-
tion of land among the rural population: countries with denserural
populations mayhavea more evendistribution of landthan countries
with sparsepopulations. The most recent data on distribution of
holdings by size is given in the worldwide censusof agriculture held
in the early1960s. Thiscovered83 countries,includingall of the larger
countries that are membersof the Bank,except Afghanistan,Bolivia,
Ecuador,Nigeriaand Romania.
The censusprovides a breakdown of distribution by size of 138.3
million holdings in the 83 countries.Thereis also a breakdownof the
distribution of land and cropland by size of holding for 64 countries
(which account for all but 9% of the land in the 83 countriescovered
in the census).Table 1:6 combines the two sets of information to
give an indication of the distribution of land and cropland by size
of holding. It showsthat:
1. About 53.9 million holdings, or 39% of the total number, are
under 1 hectarein size. If the pattern in the 83 countries is the same
as in the 64 countriesfor which there are data on distribution of size
and distribution of land, then theseholdingsoccupy1.1% of the land
areaand 3.4% of the cropland.
2. About 109 million holdings, or 78.8% of the total number, are
lessthan 5 hectaresin size. Basedon the sameassumptionas above,
theseholdings accountfor approximately6.8% of the total land area
and 20.7% of the cropland.

55
 Annex1
Table 1:6
Distributionof Holdingsby Size and Percentage
of Total Holdings:Distributionof Holdings
by Percentageof Landand Cropland
Numberof holdings All farmland Cropland
Sizedistribution Percentage in holding in holding
(hectares) (millions) distribution (%) (5Y)

Under1 53.90 38.90 1.10 3.40

1- 2 26.55 19.20 1.70 5.30
2- 5 28.73 20.70 4.00 12.00
5- 10 13.24 9.60 4.20 11.50
10- 20 7.27 5.20 4.40 10.70
20- 50 4.40 3.20 5.80 11.80
50- 100 1.97 1.40 5.80 9.80
100- 200 1.40 1.00 6.60 11.00
200- 500 0.67 0.48 8.60 11.50
500-1,000 0.23 0.16 6.50 5.90
1,000andover 0.23 0.16 51.30 7.10
Total 138.59 100.00 100.00 100.00

Source:FAO.Reportonthe 1960WorldCensus
ofAgriculture,
pp. 34-36.Rome:1971.

3. One million holdingsof 200 hectaresor more representlessthan

0.8% of all holdingsin the 83 countries.In the 64 countriessurveyed,
farms of this size group account for 66% of the total land area and
nearly25% of all cropland.
These data confirm that, when viewed in the aggregate,the dis-
tribution of land and cropland is highly skewed.If the distribution of
holdingsby size in 83 countries representsa global picture,and if the
distribution of 91% of the land reflectsthe pattern of distribution of
all the land,then holdingsabove50 hectaresin size,which represent
3.2% of all holdings, account for 78.8% of the total farmland area
and 45.3% of all the cropland. That is, roughly3% of all holdings(in
the aggregate)account for slightly less than half of the arable land
and land under permanentcrops,and more than three-quartersof all
farmland.Conversely,97% of all holdingsaccount for lessthan one-
quarter of all farmland and slightly more than half of the area under
crops.
The information on distribution of holdingsby size refersto the 83
countries, both developed and developing, covered by the census.
Therewere an estimated16 million holdingsof lessthan 5 hectares
in the developedworld: 6 million in Japanand 10 million in Europe.
Thus,of 122 million holdings in the developing countries,92 million
were lessthan 5 hectaresin size; approximatelyhalf of theseholdings

56
 Annex1
were less than 1 hectareand the remainderwere between 1 and 5
hectaresin size.
It is safeto conclude that well in excessof 100million holdingsare
less than 5 hectaresin size in the developing world at the present
time. This conclusionis derived asfollows: The 1960censusindicated
that there were approximately92 million smallholdersin developing
countries, excluding those in Nigeria, Afghanistan, Ecuador and
Bolivia. Together,at the time of the census,these countries had an
agricultural population estimatedto be close to 50 million people,
or 10 million families, most of whom were farming on units of less
than 5 hectaresin size.Thus, it is highly likely that closeto 100million
holdings of less than 5 hectaresexistedin 1960. Between1960 and
1970, the agricultural population in the developing countries in-
creasedby a reported 190 million persons,or by more than an esti-
mated 35 million farm families. Preliminaryindications are that the
fragmentationof holdings hasincreasedin manyof the more densely
populated countries as well as in countrieswhere the distribution of
land is skewed. Consequently,it is safe to assumethat the census
forthcoming in the 1970swill reveal that there are well in excessof
100 million smailholdersin the developingworld; in all probability,
more than half of their holdingsare lessthan 1 hectarein size.
The1960censusdata alsoprovided information on holdingsby size
and land areafor different regionsand countries.The most compre-
hensiveregional and national analysisfor the 83 countries dealswith
holdings of 1 hectareor more in size and pertains to 84.4 million
holdingscovering2,242million hectares.Obviously, this is not a com-
plete coverage,since it excludes holdings of less than 1 hectare.
However, it does provide an insight into the patternsof distribution
of holdings within the major regions.The resultsare summarizedin
Table 1:7.
Table 1:7
Distributionof HoldingsaboveOne Hectare,by Size and Area
1-5hectares 5-50hectares 50 hectares
% holdings % area % holdings % area % holdings % area

Europe 50.0 13.0 47.4 52.3 2.4 34.7

Northand
Central
America 23.4 0.5 39.4 8.0 37.2 91.5
SouthAmerica 36.4 1.0 45.5 8.5 17.8 90.5
Asia 78.2 40.7 21.6 50.2 0.2 9.1
Africa 73.2 3.7 23.7 6.3 3.1 90.0
Oceania 5.5 - 27.7 0.5 66.0 99.5
Source:FAO.Reporton
the 1960World
Census
of Agricalture.
Rome:1971.

57
 Annex 1

The analysisindicates the vast differences in patterns of land-

holding and land distribution between Asia and the other regions.
The contrast between Asia and the Americas is highlighted by the
fact that 78% of the holdings larger than 1 hectare in Asia are less
than 5 hectaresin size and occupy 40.7% of the land. The 36.4%
of holdings in South America and 23.4% in North and Central
America that are less than 5 hectaresin size occupy only 1% and
0.5%, respectively,of the area under farms. Only 9% of the area in
Asia is in holdings of more than 50 hectares.As much as 34.7% in
Europe, and more than 90% in North and Central America, South
America and Oceania,is in farmsof more than 50 hectaresin size.
The data for Africa, as presentedin the census,are misleading.This
is becausecoverageof that continent in the 1960 censuswas poor,
with the data on the distribution of holdings by size and acreagefor
the 18 countries surveyedheavily weighted by the results in South
Africa and Southern Rhodesia,while the sampling in Zambia was
confined to Europeanholdingsand in Tanzaniato commercialhold-
ings. If these are excluded from the sample,then the land held by
smallholdersowning under 5 hectaresis much more than 50% of all
land.
The analysisof the distribution of holdings by size on a regional
basispoints to the highly skewed distribution in the Americas; the
pattern of holdings in the eight major countries in LatinAmerica, as
shown in Table 1:8, helps explainthis. The information confirmsthat
a very high proportion of all land-ranging from 86% to 97.5%-
in the eight countries is in holdingsof more than 50 hectaresin size.
At the other end of the spectrum, only 5% of the land in the eight

Table 1:8
Distribution of Holdings above One Hectare,
by Size and Area, in Selected South American Countries
1-5hectares 5-50hectares 50 hectares
%holdings % area % holdings % area % holdings % area

Argentina 14.9 0.1 38.5 2.4 46.6 97.5

Brazil 28.1 1.0 52.6 12.8 20.3 86.2
Chile 37.7 0.7 30.3 5.2 32.0 94.1
Colombia 50.3 4.1 40.6 10.1 9.1 85.8
Paraguay 43.5 1.1 51.0 6.6 6.5 92.3
Peru 73.8 4.2 22.9 8.0 3.3 87.8
Uruguay 14.7 0.2 49.2 4.6 36.1 95.2
Venezuela 36.3 1.3 42.9 6.7 20.8 92.0

Source:FAO.Reportor the 1960WorldCensus

ofAgriculture.Rome:1971.

58
 Annex 1

countries is in holdings of less than 5 hectares(even though these

holdingsconstitutebetween 14% and 74% of all holdings).
A further partial measureof concentrationof holdings is given by
the Gini coefficient-an index of concentration based on the de-
parture of an existing pattern of holdings from an even distribution,
as revealedby a Lorenzcurve.The Gini coefficient hasbeenestimated
for 30 countries which have been grouped into three categories,as
shown in Table 1:9. As can be seen,the Gini coefficient indicatesa
high concentration in six South American countries included in the
sample.On the other hand, countries such as the Republicof China
(Taiwan),Canada,Japanand Swedenhave a low concentration of
holdings. Clearly,the distribution of holdingsby sizevarieswidely in
different parts of the world. The most skeweddistribution appearsto
be in LatinAmerica where the densityof population is relatively low
in rural areas.At the sametime, the distribution of land appearsto be
much lessskewedin many areaswith a very high densityof popula-
tion, notably Asia and Europe.It is of special interestthat two of the
countrieswith a high densityof population and very little concentra-
tion of landholdingsare Japanand Taiwan.
The distribution of land by size of holdings is "a geographical
phenomenon" and must be interpreted with caution in a socio-
economic context. It may indicate little about the international dis-
tribution of wealth or income-5 hectaresof irrigated land in Japan
would certainly yield an income well in excessof that yielded by
Table 1:9
Concentration of Land Ownership
in Selected Countries
High concentration Medium concentration Low concentration

Argentina Austria Belgium

Brazil Egypt,ArabRepublicof Canada
Colombia India China,Republicof
Iraq Iran Denmark
Peru Ireland Germany,FederalRepublicof
Spain Italy Greece
Uruguay Netherlands Japan
Venezuela Norway Philippines
Pakistan Sweden
Turkey Yugoslavia
United Kingdom
UnitedStates

Sources: FAO. Land-Tenure:WorldAgriculturalStructure,Study No. 2. Rome: 1961. Other data provided by FAO.
US Departmentof Agriculture, Economic ResearchService: ChangesinAgricultutrein726Developing
Nations,1948to 1963,
p. 36. Washington: 1965.

59
 Annex1
100,000acresin parts of Northern Australia.Similarly, within coun-
tries, the pattern of distribution of land maynot reflectthe prevailing
pattern of distribution of wealth or the socioeconomicconditions
-2 hectaresof irrigated land in the MedjerdaValley of Tunisia, pro-
ducing tomatoes,yield a far greater income than do 1,000 hectares
of land usedfor sharecroppingin the semiaridparts of Tunisia'scen-
tral area.
The caveatson quality of land and ecologicalconditions governing
land-use patterns must be borne in mind. The evidence presented
here (andelsewhere)indicates,however,that mostof the agricultural
land and cropland is concentratedin a relativelyfew holdings. It also
indicatesthat the greatestskewnessin distribution is in the Americas,
and that this skewnessis by no meansconfined to LatinAmerica.

Tenantsand FarmLaborers
The distribution of holdingsby size and population densitiesgives
no indication of the statusof thosewho hold the land or the numbers
of the landless.Only limited data on theseare available.Table 1:10
givessome information on the number of rentersand sharecroppers
in 15 countries, and the percentageof farms and areasof farmland
they occupy. Table 1:11 indicates the number of landless farm
workers in 12 countries.
This limited sample indicates that renting and sharecroppingare
widespread in all the major regions of the world. In such countries
as the Republic of Viet-Nam, Iran and Egypt, more than two-thirds
of the farms, occupying much more than half of the land, are farmed
by tenants or sharecroppers.However, in other countries, such as
Guatemalaand Tunisia, this is true of less than one-quarter of the
farms. All in all, in the 15 countries,out of 82 million holdings, close
to 29 million are worked by rentersand sharecroppers.
Renting or sharecroppingof land is a common practice in both
developedand developing countries.In someparts of the world, the
rights of those who rent land are protected by law or custom, and
rentersenjoy the sameworking conditions asownersof land. In other
areas,however, rentersand sharecroppersare in a verytenuous posi-
tion when it comes to negotiating arrangementswith the landlord,
and they commonly give as muchas half their output in return for the
useof land and servicesprovided by him.
The conditions that govern rental agreementsand crop-sharing
arrangementsdiffer throughout the world. In mostdeveloping coun-
tries, where tenancyis widespread,there is heavydependenceon the
landlord-usually an absenteelandowner-for the provision of pur-

60
 Annex 1
Table 1:10
Tenancyand Sharecropping
in SelectedCountries(l)
Renting and sharecropping
as percentage of total
Number of Number of renters
2 2
farms( ) Farmland and sharecroppers( )
-(O (%) (000)

Asia
India 27.3 n.a. 13,350
Indonesia 35.9 25.9 4,392
3
Malaysia(
) 31.2 15.7 141
4
Pakistan(
) 43.4 57.0 5,271
Philippines 54.3 40.4 1,176
Viet-Nam, Republic
of 70.3 70.0 1,334
Total 33.0 45.7(5) 25,664
MiddleEastandNorthAfrica
Egypt 62.1 57.2 1,020
Iran 66.7 73.4 1,253
Tunisia 23.3 32.0 76
Total 61.1 62.6 2,349
Latin AmericaandCaribbean
Chile 49.3 24.4 128
Colombia 31.5 13.5 381
DominicanRepublic 28.9 n.a. 129
Guatemala 22.4 16.6 93
Nicaragua 26.3 n.a. 27
TrinidadandTobago 49.5 32.8 18
Total 31.4 19.2(5) 776
(a) Data refer to latest available year in 1960s and, therefore, do not reflect land reform action on the one hand and
changes in the work force on the other.
(a) Includes holdings operated under more than one tenure form (21.81).
(1) 1960 estimates are for former Federation of Malaya.
(4) Includes both Pakistan and Bangladesh.
(5) Dominican Republic, India and Nicaragua are excluded, due to lack of data.

Source:FAO.Reportoethe 1968WorldCensus
ofAgricolture.Vol.5, pp. 92-97.Rome:1971.

chased inputs. Another widespreadcharacteristicis the absenceof

written registeredagreementsgoverning the conditions of tenancy
and the rights of tenants (eventhough there may be lawsstipulating
what these should be). Tenantsand sharecropperstypically operate
under conditions of great insecurity and are in a weak bargaining
position vis-a-visthe landlord. Frequently,the tenantsare among the
lowest income groups in agriculture. The insecurity of tenants has
been highlighted by their displacementon short notice when tech-
nological change has made it more profitable for landowners to
mechanizetheir operations-as hashappenedin Ethiopia,India and
Pakistan.

61
 Annex1
LandlessWorkers
The number of landless-farmworkers in developing countries is
increasing.Approximately100 million personsare farmwage workers

Table 1:11
Landless Farm Workers in Selected Countries(l)
Landlessworkers Active agricultural
as % of active population as % of
Number of land- population in total active popu-
less workers agriculture lation

Asia
2
India(
) 47,300 32 68
Indonesia 5,673 20 70
3
Pakistan() 8,013 29 70
Total 60,986 30 68
MiddleEastandNorthAfrica
Algeria 1,099 60 56
ArabRepublic
Egypt, of 1,865 38 55
Iran 903 25 46
Morocco 484 19 61
Tunisia 210 20 46
Total 4,561 33 58
LatinAmericaandCaribbean
Argentina 694 51 15
Brazil 3,237 26 44
Chile(1971) 378 66 28
Colombia 1,158 42 45
CostaRica 122 53 45
DominicanRepublic 179 25 61
Ecuador 391 39 54
Honduras 138 27 67
Jamaica 72 41 27
Mexico
(1970) 2,499 49 39
Nicaragua
(1971) 101 43 47
Peru 557 30 46
Uruguay 99 55 17
Venezuela 287 33 26
Total 9,912 35 39
(1) Except for India, data presented here are estimatedfrom [LO, YearBookof LabourStatistics 1871,pp.43-294, and
1972. pp. 44-301. Unless otherwise indicated, data refer to latest year available in 1960s and, thus, do not reflect
recent reform actions on the one hand and changes in the work force, on the other.
(2)Agricultural laborers as shown in India: Ministry of Agriculture, Directorate of Economics and Statistics. indian
Brief (I Ith ed., 1971), p. 14.
Agricurltureinn
(3) Includes population now belonging to Bangladesh.

62
 Annex1
(including family membersand headsof familieswith verysmall land-
holdings) in the 22 countries for which data are provided in Table
1:11. This figure includes an estimated47 million in India alone-
about 32% of the active population in agriculture. There are about
10 million suchworkers in LatinAmerica.Evenin Argentinaand Uru-
guay (with only 15% of the active population depending on agricul-
ture), more than half of the workers are essentiallylandless.In the
remainingcountriesof the region, the proportion rangesfrom a mini-
mum of about one-fourth in Brazil and Hondurasto a maximum of
approximatelytwo-thirds in Chile,
Almost no reliable estimatesexist of the number of unemployed
in rural areas.It is usuallyassumedthat the labor force subsistsoff a
holding and joins in some arrangementwith the extended family
whereby it shareswork and output. The emergenceof a landless
wage-earningclassconfirms that a growing rural labor force hasto
rely on work outside the traditional sectors for its livelihood. This
group is increasingin size,and the provisionof employmentfor what
is alreadya large rural proletariat may well be one of the greatest
challengesfacing national governmentsin the future.
There is a vast amount of underemploymentin the rural areasof
most countries of the world. The nature of this phenomenon has
been discussedelsewhere.At this juncture, it should be pointed out
that the redistribution of idle land can provide added employment,
but that the prospect is limited for redistribution of land providing
full employment for all the presentand prospectivepopulations in
the rural areas of densely populated countries. Structural changes
within agriculture can help alleviate underemploymentand open
unemployment,but the problemsof reducingnationwide unemploy-
ment haveto be seenin a national ratherthan a sectoral context.

63
 Annex2
EXPERIENCESWITH
LAND REFORM
The following summariesillustrate selectedcountry experiencein
land reform over the lastthree decades.Their inclusionin this paper
should not be taken as indicative of Bankjudgment on what doesor
does not constitute land reform, nor should the statementsbe re-
garded as definitive. Land reform is a complex process in which
severalsocioeconomicvariablesare changedmore or lesssimultane-
ously.In mostcases,the evidenceis inadequateto allow identification
of causalrelationshipsbetween reformmeasureson the one hand and
production, income and social effectson the other, even though it is
often feasible to trace correlations,such as that between land dis-
tribution and a rise in productivity.

Republicof China
Taiwan'sland reform program was implementedin three steps.A
reduction of rents, in 1949,was followed by the saleof public lands.
A land-to-the-tiller programcompletedthe reform in 1953.The pro-
portion of cultivated land under tenancy leaseswas reduced from
41% to 16%, while the proportion of farm families owning all land
under their cultivation increasedfrom 33% to 59%. On the land
remainingunder tenancy cultivation, written and secureleaseswere
arrangedat much reducedrental rates.
Following the reform,the productivity of agriculturehasincreased,
income distribution has become more even, and rural and social
stability havebeenenhanced.Landproductivity is higheston holdings
below 0.5 hectare.The shareof total agriculturalincome that is con-
sumed hasincreasedonly moderately,leavingintact enoughincome
to achievea fairly high agricultural savingsrate.
The smooth implementation of the reform programin Taiwan was
due to a stable sociopolitical climate and the many complementary
developmentmeasurestakenbefore and during the reform.Theexist-
ence of a thorough cadastralsurvey,good agricultural researchand
extensionservices,a vast expansionof publicly sponsoredfarm credit
during the reform period, and a graduallyincreasinginvolvementof
tenant farmers in the administration of the program,all contributed
to the success.
Republicof Korea
Land reform in SouthKoreaafter the SecondWorld War consisted
of: (1) a reduction of farm rents from 40-60% of production to 33%

64
 Annex2
in 1945; (2) a redistribution, in 1948, of Japaneseproperty confis-
cated by the military authorities; and (3) a redistribution between
1950 and 1953 of land in excessof a ceiling of 3 hectareson Korean
holdings.The terms of salewere similarly generoustoward the buyer
in both cases.Some 1.4 million acres (25% of the total farmland)
were distributed to 1.6 million farmers (approximately70% of all
farmers).
It.hasbeen estimatedthat, before the reform, 19% of the farmers
owned 90% of the land and more than 50% of the farmerswere land-
less tenants.Afterward, 69% of the farmers owned all the land on
which they worked and 24% were part-owners,while only 7% were
tenants.Considerablesociopolitical stability has been achieved,to-
getherwith income redistribution in favor of the poorer ruralfamilies.
Yields did not fall as a consequenceof the reform; by the 1960s,
yields had far surpassedprereformlevels.Laborproductivity and rural
employment increased.But the small size of most farms has now
becomea constrainton farm income.
Japan
The first Japaneseland reform program, in 1868, laid the ground-
work for Japan'ssocial and economic transformation.The peasantry
wasfreed from bondage,the power of the feudal lordsto collect taxes
from landownerswas broken, and private landownershipwas rein-
forced for the purposeof cash taxation by the central government.
Supplementaryprogramsfor infrastructureimprovement,training and
extension,credit services,and promotion of farm chemicalsand new
crop varietieswere pushedon a large scale.Labor intensityand land
productivity rose quickly, with the result that the agricultural sector
could provide savings,cheapfood and surpluslabor to the industrial
sector. The first reform did little, however, to distribute property
ownership or reduce income inequality-rather it strengthenedthe
landownerclass.
Subsequentto the first reform, the tenancyproblem grewgradually
worse. Largenumbersof smallholderslost their property in the agri-
cultural depressionat the turn of the century,partly becauseof heavy
land taxes.In the late 1940s, a secondland reform programwas exe-
cuted. Owners had to sell all land in excessof about one hectareto
the governmentat confiscatoryprices.Theformer tenantsweregiven
property rights at an extremely low real cost, which resulted in a
thorough restructuringof rural society.
The second reform resulted in greater equity, and may also have
removed a constraint on the growth of Japaneseagriculture. The
economiceffects were not as enormousas thoseassociatedwith the

65
 Annex2
first reform. Land productivity did increase after 1947, but some
observersregardthis asessentiallya continuationof a long-term trend
(1895-1939) startedby the first reform.
The second reform worsened,however,the problems of fragmen-
tation and undersizedfarms.At the time of the reform, the tenancy
problem had already been relieved through a reduction of excess
rural population by the war and absorption into industry. The land-
lords who were forced to sell excessproperty were mostly small-
holders themselves.Two-thirds of the owners were required to sell
lessthan one hectareand only 6% more than five hectares.Although
the reform increasedincome equality among farmers, it hampered
equalizationof rural and urban incomes.Part-timework outside the
farm is an outlet, but the farmersconcernedare often limited to low-
skilled work. Rural incomes have, therefore, lagged behind, price
supports notwithstanding.An attempt to create larger farming units
through cooperativeshashad little effect. Agricultural policy is now
aimedat, amongother objectives,an increaseof farm incomethrough
diversification into horticulture and animal husbandry.

India
Land reform in India, pursued since 1950-51, is largely recom-
mendedand coordinated by the Central Governmentand the Plan-
ning Commissionand executedby the individual stategovernments,
with the result that policy implementation varies widely. The four
major types of reform havebeen: (1) the abolition of the zamindari'
system;(2) tenancyreformdesignedto fix maximumrents,to improve
security of tenure and to give the right of purchaseto the tenant; (3)
ceilingson landownershipand distribution of surplus;and (4)consoli-
dation of fragmentedholdings.
By 1961, the intermediary rent and tax collectors, most impor-
tant of whom were the zamindars,had been abolished.Sincetenants
continue to pay revenuedirectly to the government,their economic
position hasnot been greatly improved. The abolition of the zamin-
dari systeminvolved 173 million acres, more than half of the area
occupied by holdings. A total of Rs.4,350 million was paid in com-
pensation,mainly in the form of bonds.
Under the tenancy reforms, 3 million tenants, subtenantsand
sharecroppershad, by 1961, acquired ownership under purchase
agreementsof 7 million acres.Securityof tenure appearsin general
to haveworsened, however. Actual rents have not come down; in
'The zamindars were revenue collectors during the Moghul period. Under the British,
they gradually turned into powerful landlords.

66
 Annex2
some statesthey have even increased.Landownershave been per-
mitted to resumeland above legal ceilingsfor personal cultivation,
which has allowed them to escapethe reforms. Unreported casual
tenancyand shareagreementshavemultiplied.
Under the ceilings legislation,approximately2 million acreshave
been taken over by the government in order to settle tenantsand
landlesslaborers.A further 4.2 million acreswere formally pledged
to the Bhoodan(gift) movement,but most of the donatedparcelsare
still in the handsof the donors.Only about 1 million acresout of all
gifted land haveactually been given to landlesslaborers.
Consolidation of land parcels has been more successfuland has
resultedin a rationalizationof holdings covering 69 million acres.It
appearsto havecontributed to a growth in productivity in the north-
ern statesof Punjab,Uttar Pradeshand Haryana.
It is well recognizedin India that the reform measuresdealingwith
securityof tenureand acreageceilingsareonly partiallyenforced,and
that manyof the statelegislaturesare not anxiousto havesuch radical
land reform.Evenif a ceiling is imposed,the landacquiredis sufficient
to give minimal holdingseither to the minifarmersor the landless-
but not both. Thereappearsto be scopefor somedistribution which
will also assistagricultural production becausethe yield per acre in
India is higher on small farms. As long as population pressurecon-
tinues, it will be unrealisticto try to abolishtenancy in the short run.
Therefore,it will be better to legalize someforms of tenancywhich
exist on a largescale,and to promote more efficient typesof tenancy
contracts.All kinds of tenants should also be registeredand given
accessto credit and inputs. A large extensionof credit at reasonable
terms, together with accessiblemarketingchannelsto small farms in
general,and particularly to tenantswith secureleases,is required.
Provisionof thesefacilities is as essentialas further land distribution
for attaining the income equity and productivity objectivesof India's
land reform, and is likely to presentfewer problems.

Iran
Iran's land reform started in 1962. Before the reform, 56% of the
holdings, covering 62% of the area under cultivation, were rented.
Tenantswere rotated annually, a practice which hampered agricul-
tural investmentand causedexploitative useof the soil. The.largest
estatesoccupied relativelymore fertile lands,and ownerswere often
absenteelandlordswho contributed little to agriculturalproduction.
Former landownerswere partly compensatedupon expropriation
by cash paymentsrangingfrom 10% to 20% of the estimatedvalue

67
 Annex2
of their holdings, with the balancepaid in bonds in annual install-
ments.The beneficiarieswere to repaythe governmentthe expropri-
ation price plus 10% to cover administrativecharges.As these pay-
mentsfell behind, the Central Bankfunded the difference.The costs
to the Governmentwere limited to thoseincurred in carryingover the
acquisition coststo the time of final reimbursement.
During the first stageof the reform, landownershipwas limited to
a maximum of one village per owner. Excessland was expropriated
and distributed to the tenants. In the second stage,the limit of one
village was reduced further to plots of 20-100 hectares(depending
on the natureand location of the land). The landlord had five options
for the area in excessof the maximum allowed to him, to wit: (1)
leasingto the tenantsfor 30 years; (2) selling to the tenants; (3) pur-
chasingthe tenants' rights; (4) dividing the land with the tenants in
the same ratio as the customary crop sharing; and (5) forming an
agricultural unit for joint operation by the owner and the tenants.
The third and final stageof the reform,which was practicallycom-
pleted in 1971, aimed at conversionof all 30-year leasesinto small-
holdings. Virtually all of Iran's 50,000villageshave undergone land
reform and more than 3 million families have received land.
Although agricultural output increasedby a total of 18% in the
first five yearsof the reforms, it is believedthat the land reform pro-
gram on balancehad adverseshort-run effects on output. It created
uncertainty which discouragedinvestment in improvements; there
was also considerableinterferencewith the normal flow of irrigation
water from streamsand storageplacesstill controlled by landlords.
The reform favored tenants and sharecroppersinsofar as it con-
ferred ownership on them or enhanced their security of tenure.
Becausethey were basedon the existingdistribution of holdings,the
reforms did not assistthose who were landless.Continuation of the
existing inequities of land distribution was regardedas one of the
costsof ensuringa speedyenactmentof the reform.
The ownershipand tenancy reforms havebeen complementedby
rural cooperatives,credit and extensionservices,and increasedsup-
ply of quality seedsand fertilizers. Many measureswere set up in a
somewhat improvised fashion. The early accomplishmentsof the
credit program were striking; total lending by the Agricultural Bank
tripled between 1960and 1965,but this growth leveledoff after 1966.

Morocco
The Moroccan Government has undertakena series of measures
aimed at land reform since independencein 1956. The objective of
these measuresis to facilitate an increasein agricultural production

68
 Annex2
and to improve the distribution of rural incomes.Legislationpassed
in 1962,1966 and 1972 provides for land consolidation and distribu-
tion of land to smallholdersand landless families throughout the
country. The Agricultural Investment Code, published in 1969, is
aimed at facilitating the developmentof irrigated agriculturein well-
defined developmentzones. It providesfor the restrictionof inheri-
tance rights to limit fragmentation, an improvement in the tenure
position of membersof traditional collectives,and the adoption of
modern cultivation techniques.
Land distribution is so far basedmainly on former foreign-owned
land,althoughsomeother state-ownedland and traditional collective
land is involved. At the time of independencein 1956,about 900,000
hectareswere foreign-owned; of this area, about 300,000hectares
were sold privatelyto Moroccans,mainly before 1963,when legisla-
tion was introduced subjecting such transfersto Government ap-
proval. Thirty-one thousand hectareswhich were mainly used by
foreignersfor researchpurposeswere recoveredby 1960, and a fur-
ther 220,000hectaresof "official colonization" landswere takenover
by the Government between 1963 and 1965.
Distribution so far hasbeen limited to land underfield crops,while
land under tree crops (mainly orangegroves)remainedunder Gov-
ernment control and ownership. Distribution to smallholdersand
landlessfamilies was slow until 1967 and then gatheredmomentum
up to 1972. By the end of 1972, 181,000hectares(3% of the culti-
vated area)had been distributed to over 11,000families. However,
the impact of land distribution alone on the problem of rural poverty
hasbeen small; the number of beneficiariesso far is only about 1%
of farm families with lessthan 2 hectares.
Through the establishmentof cooperatives,intensified extension
supportand the provision of modern inputs, the beneficiariesof land
reform have generally quickly achieved high yields and acceptable
incomes.Land consolidation hasalso been successfuland hasso far
benefited almost 200,000hectares.The main constrainton the pro-
gram hasbeen the unavoidablecomplexity of supervisingits imple-
mentation consideringthe Government'smanpower resources.The
Government'smain priority now is to accelerateland distribution,
while maintaining high technical standardsof managementon the
distributed land. Remainingforeign-ownedland,amountingto about
370,000hectares,was recovered by the Government in 1973. The
target for the third Five-YearPlanis to distribute 395,000hectaresof
land under field crops, mainly formerly foreign-owned, between
1974 and 1977, and to seeka suitable formula for distributing land
under tree crops.The achievementof the distribution target for land

69
 Annex2
under field crops alone would, by the end of the plan, enable the
program to cover 9% of cultivated areaand 5% of farm familieswith
lessthan 2 hectares.

Yugoslavia
The first land reform in Yugoslaviawas undertakenin 1919. In the
south and west, bondage was abolished, and the tenants of the
Turkish landownersreceivedownership rights. In the north, the size
of the large estateswas reduced, but the former landownerswere
allowed to retain ratherlargeholdings.The implementationtook two
decades,and resulted in a transfer of ownership of almost 25% of
the farmlandto more than 33% of the peasants.
The second land reform started in 1945,when all large estates,all
land in excessof 25-35 hectaresper farm, and the farm property of
Germansand other aliens,were expropriated.Half of the seizedland
was distributed to the poor and landless,while the other half was
retainedas state property. The stateand collective farms createdin
the late 1940salong Soviet lines expandedto approximately25% of
the total cropland. Collective farms were allowed to disband after
1952,however,and by 1956accountedfor only about 10% of all land
under cultivation.
Aside from the socialistsector,the privatesectorof individualown-
erswho cultivate their own land remainsimportant, and vasttractsof
mountain pasturesarestill undertraditional,collectiveforms of usage.
In 1953,a ceiling of 10 hectaresof arable land or its equivalent was
imposed on private holdings.The averageholding in the private sec-
tor is now only 3.9 hectares.The socialistsector includesstatefarms,
producer cooperatives and general cooperatives. The kombinats,
which resemblethe worker-managedindustrialfirms, form the largest
and fastest-growingsocialistelement,whereasthe producer coopera-
tives havedeclined.The generalcooperativesare mainly associations
for joint input purchases,equipment useand output sales,and have
expandedto about40% of all smallholdings.
The socialistsectoris reportedlythe mrst productive.Thisis related
to the location of holdings on the better soils and its priority treat-
ment in the allocation of inputs such as fertilizers, machineryand
expertise. However, the bulk of agricultural output still originates
from the large group of small farms, consistingof both the coopera-
tives and the farms outside the socialistsector.The reforms have re-
sulted in a sizable redistribution of rural income and an increasein
peasantparticipation in rural decision making, particularly sincethe
mid-1950s.

70
 Annex2
Kenya
Land reform was initiated in Kenyaby the colonial administration
in 1954 and expanded by the Government after independencein
1963. The reform aimedat solvingseveralproblemsat the sametime.
Theseincluded: (1) adjudicationand consolidationof holdingsunder
cultivation by African farmers; (2) resettlementof African farmerson
the large farmspreviouslyowned by Europeans;(3)promotion of cash
cropping and dairying, and increasedproduction for the market;and
(4) diversificationof export output. More than 1 million acresof land
formerly cultivated by Europeanswere opened up to Kenyansmall-
holders,and the rightsto about 7 million acreswere adjudicatedand
consolidated.
The implementation and results of the reforms have been quite
successful,notwithstanding political friction and a lack of qualified
personnel.An activeextensionprogramhasenabled smallholdersto
increase the production of coffee, pyrethrum, maize, wheat, dairy
products and beef. The economic benefits of the adjudication and
consolidation of holdings seemto have been greater than those of
resettlementon largefarms.Socially,the reformshavecreateda class
of prosperoussmallholders.In particular, those that were already
relatively well-to-do have profited, while the poorest smallholders
and nomadshave benefited much less from the reform. It was esti-
mated in 1973 that approximately25% of all smallholdingswere less
than one hectareand about 50% less than two hectares,occupying
altogether lessthan 4% of total arable land. The landlessamount to
approximately16% of the rural population.
Mexico
Having its roots in the revolution of 1910-15, the agrarianreform
in Mexico createdvillage groups (ejidos)with usufruct rightsto land.
Most of the ejidos wereformed in the late 1930s and havebeen oper-
ated on an individual rather than collective basisby the ejidatarios.
Closeto 90 million hectareshavebeen distributed between1915 and
1972 to about three million ejidatarios.Theseprimary beneficiariesof
the reform represented53% of all farmersand 26% of the rural labor
force. Somethree million landlessrural workers remainand, despite
the considerableconcentration of ownership that persistsin the pri-
vate sector,1976 hasbeen plannedasa terminalyear for land reform.
Total production by the ejidos grew very slowly during the first
decadeof their establishment.Sincethen, the ejidos haveincreased
output about as fast as hasthe private sector. Incomesof the ejida-
tarios arealmost certainlybetter than would havebeen the casewith-
out reform, but substantial regional differences persist in natural

71
 Annex2
resourceendowment and in the extent of public investmentin com-
plementary infrastructure.More such investment and a mechanism
for selectiveconsolidation of small farms will be required to ensure
that the impact of the reform is maximized.
Following the land redistribution during the 1930s,the concentra-
tion of landownershipincreasedagainbetween1940and 1960.Since
then, the concentration may have fallen back as a result of the dis-
tribution of another 35 million hectares during the last decade.
Rural income distribution is still skewed. In 1967-68,50% of the
farmers earnedonly 20% of all farm income (including personalin-
come from sourcesother than agriculture).Among ejidatarios,how-
ever, income was more evenly distributed. While the top 20% of
private farmersreceived60% of all privatefarm income,the top 20%
of the ejidatarios accountedfor only 45% of all ejido income.

Peru
Betweenthe start of land reform in 1963 and 1972, a total of 4.7
million hectareshas been expropriated.Over 100,000families have
been settled on 2.8 million hectaresof this area. Expropriatedlands
that havenot yet been resettledcontinue to be operatedunder direct
governmentsupervisionuntil a cooperativeor SAIS(SociedadAgri-
cola de InteresSocial) farm organizationhasbeen formed, to which
the land title is then transferred. Despite the priority given by the
government,implementation is well behind schedule.The target for
the current Five-YearPlanis to expropriate26,200farm units contain-
ing 12 million hectares,and to redistribute theseto 500,000families.
In 1972,about three-quartersof the target area still remainedto be
expropriatedand reallocatedbefore the end of 1975.
The agrarianreform law of 1964concentratedon redistribution of
inefficiently managedlatifundia (large landed estates)in the Sierra.
Well managedproductive units were exempted.The more funda-
mental reform law of 1969was the basisfor the expropriation of the
large,productiveand profitable sugarcomplexesof the north coast.A
limit was establishedon the size of holcdings(150 hectareson the
coast). The government bonds given to the former owners can be,
used for investmentin industry to supplementtheir other resources.
Four different categoriesof farm organizationscan receive redis-
tributed land, but the bulk hasbeen placed in the handsof worker-
owned cooperatives.Only a small number of individual farms has
been assignedto former tenants,while in a few casesland hasbeen
added to the holdings of Indian communities.
TheSAISisa unique form of farm organization,and is the basicunit
of agriculturalreform in the Sierra.TheSAISrepresentsan attemptto

72
 Annex2
solvethe problem of providing agricultural and social development
opportunities to the membersof the traditional Indian communities
without jeopardizing the relatively high production and economies
of scaleattainableon expropriated haciendas.Haciendaproduction
is almost entirely basedon extensivegrazing of mountain pastures,
and early experiencesof land distribution in the Sierra indicated a
high risk to production if haciendaswere taken over as community
land or subdivided into small sheep ranches.
* In anyattemptto meet socialneedsthrough redistributinglandand
income in the Sierra,therefore, the governmentis faced with prob-
lems of maintaining or raising productivity levels attainable only
through exploitation of scale economies.The SAIS,the proposed
solution to this dilemma, accountedin 1972 for 10% of the families
benefiting from the agrarianreform program. It can be regardedas a
second-degreecooperativewhosemembersaresocialbodies instead
of individuals.Membershipof eachSAISunit consistsof the coopera-
tive of the production unit and of the communities surrounding it.
Eachgroup contributesto the capitalof the enterpriseon the basisof
resources,population and economic potential; the share of each
group is determined by the land reform agency.Managementof the
SAISis in the handsof professionalemployees.Profitsare allocatedto
each membercommunity in relation to its sharein the SAIS,and are
to be used in community development projects involving schools,
roads, power reticulation and housing. In this manner,surplus man-
power is given employment, and the rather meagerprofits can be
usedin developingbadly neededphysicalinfrastructure.
The debt assumedby eachSAISunit is to be repaid from profits in
20 years following a five-year grace period. Debt repaymentmay
becomean onerousburden on those units whose profit potential is
limited by their physicalcapacityto expandlivestocknumbersand by
the need to employ high-quality technical services.Legally,the full
market value of expropriatedlivestock hasto be paid in cashwhile
fixed capital is to be paid for largely in agrarianbonds.
The land reform programalone will not be able to solvethe rural
,unemployment problem. Evenif the optimistic targets for 1975 are
met, employmentopportunities in agriculturewill increaseonly from
1.32million to 1.6 million, while the number seekingwork in agricul-
ture will rise from 1.9 million to 2.1 million. Nearly800,000families
with insufficient land to provide adequatesubsistenceare eligible to
benefit through the land reform program. Evenif all land which can
be expropriatedis redistributed,about 500,000families,mostly in the
Sierra,will still lack a minimum subsistencelandholding. However,
agrarianreform is providing the basisfor socialand economicchange.

73
f
World BankOffices
Headquarters:1818 H Street, N.W., Washington,D.C. 20433, U.S.A.
New York Offices: c/o United Nations, Room 2245, SecretariatBuildings,
New York, N.Y. 10017, U.S.A.
120 Broadway (15th Floor),New York, N.Y. 10005, U.S.A.
EuropeanOffice: World Bank,66, avenue d'14na,75116 Paris,France
LondonOffice: World Bank, New Zealand House (15th Floor),
Haymarket,London, SWi Y4TE,England
Tokyo Office: World Bank,KokusaiBuilding, 1-1 Marunouchi3-chome,
Chiyoda-ku, Tokyo 100, Japan
EasternAfrica: World Bank RegionalMission, ExtelcomsHouse,
Haile SelassieAvenue, Nairobi, Kenya; mailing address-P.O. Box 30577
Western Africa: World BankRegionalMission, Immeuble Shell,
64, avenue Lamblin,Abidjan, Ivory Coast; mailing address-B.P. 1850
Afghanistan:World BankResident Mission,P.O. Box 211, Kabul, Afghanistan
Bangladesh:World Bank ResidentMission,BangladeshBank Building
(4th Floor),Motijheel Commercial Area, G.P.O.Box 97, Dacca,Bangladesh
Cameroon:World BankResidentMission, Immeuble Concorde,angle avenue
El Hadj Ahmadou Ahidjo et avenueJ.F.Kennedy,Yaounde',Cameroon;
mailing address-B.P. 1128
Colombia:Resident MissionBanco Mundial, EdificioAseguradoradel Valle,
Carrera10 No. 24-55, Piso 17, Bogota D.E.,Colombia
Ethiopia:World BankResidentMission, I.B.T.E.New Telecommunications
Building (FirstFloor), Churchill Road, Addis Ababa, Ethiopia;
mailing address-IBRD Mission, P.O. Box 5515
Ghana: World Bank ResidentMission,c/o Royal GuardianExchangeAssurance
Building, Head Office, High Street (5th Floor),Accra, Ghana;
mailing address-P.O. BoxM27
India: World BankResidentMission, 53 Lodi Estate,New Delhi 3, India;
mailing address-P.O. Box416
Indonesia:World BankResidentStaff, JalanWahid Hasyim 100,
Jakarta,Indonesia;mailing address-P.O. Box 324/JKT
Nepal: World Bank (IBRD)ResidentMission, R.N.A.C.Building (FirstFloor),
Kathmandu,Nepal; mailing address-P.O. Box 798
Nigeria: World BankResidentMission, 30 Macarthy Street, Lagos,Nigeria;
mailing address-P.O. Box127
Pakistan:World BankResidentMission, P.O. Box 1025, Islamabad,Pakistan
SaudiArabia: World Bank ResidentMission, ElectricCompany Street, Riyadh,
Saudi Arabia; mailing address-P.O. Box5900
Sudan: World BankResidentMission, 28 Block 2H, BaladiaStreet,
Khartoum,Sudan; mailing address-P.O. Box 2211
Tanzania:World Bank ResidentMission, N.lC. Building (7th Floor, B),
Dar es Salaam,Tanzania;mailing address-P.O. Box 2054
Thailand: World Bank RegionalMission,Udom Vidhya Building,
956 RamaIV Road, SalaDaengh,Bangkok5, Thailand
Upper Volta: World BankResidentMission, avenue MonseigneurThevenoud,
Ouagadougou,Upper Volta; mailing address-B.P. 622
Venezuela:World BankResidentMission, Centro Andres Bello,Avenida
Andres Bello,113-E,Mariperez,Caracas,Venezuela
Zaire: World Bank ResidentMission, Building UZB, avenuedes Aviateurs,
Kinshasa1, Republic of Zaire; mailing address-P.O. Box 14816
Zambia: World Bank ResidentMission, Kulima Tower (13th floor),
Katunjila Road,Lusaka,Zambia; mailing address-P.O. Box 4410
 i

I
A I '
World Bank
Headquarters:
1818H Street,N.W.
Washington,D.C. 20433, U.S.A.
Telephone: (202) 393-6360
U
Cableaddress: INTBAFRAD
WASHINGTONDC